vectorcraft/docs/mcp.md

VectorCraft MCP server

vectorcraft-cli mcp runs a Model Context Protocol server on stdio (newline-delimited JSON-RPC 2.0; see Protocol for version negotiation). Agents use it to draw, inspect and look at VectorCraft documents.

It has two backends:

Mode	How	What works
Remote	vectorcraft-cli mcp --connect 127.0.0.1:7979 (the app must run with vectorcraft --control 7979)	Everything. Tool calls are forwarded over the control protocol, so you watch the app change live.
Headless	vectorcraft-cli mcp --headless	An in-process engine session with a CPU renderer. Everything except the UI-only tools (inspect_ui, type_text, open_panel, screenshot {window:true}).

With no flag, the server tries 127.0.0.1:7979 and falls back to headless. Logs go to stderr; stdout carries only protocol messages.

Install and register

From the repository root, with Rust 1.95+ and the platform's source-build prerequisites installed:

cargo install --path apps/vectorcraft-cli --locked
# POSIX shells, for Cargo's default install location:
export PATH="${CARGO_HOME:-$HOME/.cargo}/bin:$PATH"
vectorcraft-cli --help
claude mcp add vectorcraft -- vectorcraft-cli mcp
# or pin a mode:
claude mcp add vectorcraft-headless -- vectorcraft-cli mcp --headless
claude mcp add vectorcraft-app -- vectorcraft-cli mcp --connect 127.0.0.1:7979

Cargo installs a binary, not a symlink into target/, so it survives cargo clean. Re-run the install command after updating the checkout. By default the executable goes in $CARGO_HOME/bin (~/.cargo/bin when unset), or %USERPROFILE%\.cargo\bin on Windows. If you set --root, CARGO_INSTALL_ROOT, or Cargo's install.root, use that installation root's bin directory instead. Add it to the PATH inherited by the MCP client, not just an interactive shell, and restart the client after changing its environment. Installing does not start or register a server.

Other clients use the same command in their JSON config:

{"mcpServers": {"vectorcraft": {"command": "vectorcraft-cli", "args": ["mcp"]}}}

If a client reports ENOENT or "Executable not found in $PATH", verify the install with command -v vectorcraft-cli (PowerShell: Get-Command vectorcraft-cli) in the client's environment. If that environment cannot be changed, set JSON command to the absolute installed binary path, not a literal ~ or $HOME expression. Avoid symlinks into target/release: cleaning the build directory breaks them. A source build alone does not put the CLI on PATH.

For a live session, start the app first: cargo run --release -p vectorcraft -- --control 7979.

From an installed release

The release packages ship vectorcraft-cli alongside the desktop app, so no build is needed:

Install	CLI
Windows (MSI)	C:\Program Files\VectorCraft\vectorcraft-cli.exe by default (wherever you installed it otherwise), not on PATH
Linux (deb, rpm)	/usr/bin/vectorcraft-cli
macOS	the separate vectorcraft-cli-<version>-macos-<arch>.zip release asset (the .app holds only the desktop app)

# Windows, default install folder
claude mcp add vectorcraft -- "C:\Program Files\VectorCraft\vectorcraft-cli.exe" mcp
# Linux, or macOS with the CLI unzipped onto PATH
claude mcp add vectorcraft -- vectorcraft-cli mcp

Confining file access

vectorcraft-cli mcp --automation-read-root /work/project --automation-write-root /work/project/out

{"mcpServers": {"vectorcraft": {"command": "vectorcraft-cli",
  "args": ["mcp", "--automation-read-root", "/work/project", "--automation-write-root", "/work/project/out"]}}}

The two flags (PhotoCraft and the other Craft apps take the same ones, so one client config works across the suite) keep the agent to the folders you give it:

• Every file a command reads must lie inside the read root: open_file, file.place, relinking and updating links, the images an SVG links to, swatch library, graphic style library and colour-profile loads, plug-ins, file.newFromTemplate, preset imports, shortcut sets (shortcutSet.read), the folders text.findFontFiles and links.findFiles search, the scripts script.check reads and the files they include.
• Every file a command writes must lie inside the write root: save_file, export, the save and export commands, Export for Screens, Package, Export for Web, print to file, screenshot {path}, library and preset saves to a path, the copies links.relink {copy} makes.
• file.list {path} (through run_command) lists a folder's files and folders, held to the read root like any read, so an agent can see what it was given (#1163).
• Read and write access are separate. Give the same folder twice for both; a root left out grants none of its access (with only --automation-read-root, nothing can be written to disk, though export and the save commands without a path still return the bytes). Without either flag nothing changes.
• A path is made absolute against the server's working directory, then its links are followed: a file (or folder) link that leads out of a root is outside it. For a file that doesn't exist yet, its nearest existing folder's links are followed and the rest must be plain names, so .. below a folder that doesn't exist is refused, as is a link that leads nowhere (writing through it would create its target). The result is compared with the root by whole folder names: /work2/x is not inside /work. On Windows case is ignored, UNC paths (\\server\share\…) are resolved like drive-letter ones, and device names (CON, NUL, COM1…) are refused.
• A refused path is an ordinary tool error the agent can read and correct, for example automation path rejected: outside the write root /work/project/out: /etc/x.png or automation filesystem access is not granted: write authority is absent: /work/project/out/x.png. Links in an opened document that point outside the read root show as missing.
• The preferences that move folders the app reads or writes on its own (fontsFolder, pluginsFolder, recoveryFolder, templatesFolder) can only be set inside the roots. The app's own folders (preferences, Data Recovery, the User Defined library folders, VectorCraft's Fonts folder) are not confined: no agent names a path there.
• The roots must be existing folders; a missing one stops the server with an error. Each flag is given once (--flag <dir> or --flag=<dir>).

The flags confine the headless server and imply --headless; --connect with them is an error, since the server would otherwise talk to an app that isn't confined. A running app confines itself: start it with the same flags, `vectorcraft --control 7979 --automation-read-root

--automation-write-root

`, and connect as usual (see [Control protocol](control-protocol.md#confining-file-access)). The server then writes nothing itself: `screenshot {path}` asks the app to write the file.

The checks hold for every thread of the server process. A path is checked, then opened: another program that swaps a link in between can still win that race, so the roots keep an agent to the files it was given rather than guarding against other software on the computer.

Protocol

Newline-delimited JSON-RPC 2.0 on stdio. The revision is 2025-06-18; 2025-03-26 and 2024-11-05 are accepted too. Every request and notification must carry "jsonrpc": "2.0": anything else is refused with -32600 (requests) or ignored (notifications, which have no id to address a reply to) before it is dispatched. When a tool call reaches a running app, it is one control-channel request, and the app echoes its id in the reply: a reply with any other id is rejected, the connection dropped, and the next call starts a fresh one — the call reports uncertainty like any reply that never arrived whole. A malformed JSON reply also drops the connection and reports that uncertainty; the failed operation is never replayed automatically. The app's refusals carry no id (a connection past its 16, a line it would not read): the call reports the app's reason and says nothing was run. Before each call the server checks that the app still holds the connection and opens a new one when it doesn't (the app closes a connection left idle for 5 minutes, and a restarted app is a new one), so a request is never written to a closed socket; when the app can't be reached, the call says that nothing was sent. Clients requesting 2026-07-28 through initialize, or through a request's _meta["io.modelcontextprotocol/protocolVersion"], receive resultType: "complete", ttlMs and cacheScope: "private" on tool/resource/template/prompt lists and resource/prompt reads. List TTL is 600000 ms; live reads use zero. Older clients keep the original response shapes. This does not add streaming or input requests — see Not implemented.

One message is one line of at most 16 MiB: a longer line is answered with -32600 and skipped, and the server keeps serving what follows. A batch holds at most 64 messages. command_batch holds at most 256 steps, and the step results it keeps (for later steps' "$N.path" references and the reply) total at most 8 MiB — past that the batch stops with the reason, closing its atomic undo group as usual. Replies from a connected app are refused past 256 MiB and the connection dropped; the call reports uncertainty like any reply that never arrived whole.

initialize advertises:

Capability	What it covers
tools	The tools below
resources	Three fixed resources and four templates
prompts	Five ready-made workflows
completions	completion/complete for prompt arguments and template variables
logging	logging/setLevel and notifications/message

Prompts

prompts/list offers five workflows; prompts/get returns one templated user message with its arguments filled in. A missing required argument is -32602.

Prompt	Arguments (required first)
poster	brief, palette, text
icon-set	subject, count, detail
recolor	palette, method
trace-and-style	path, preset, effect
export-set	formats, directory, scale

echo '{"jsonrpc":"2.0","id":1,"method":"prompts/get","params":{"name":"poster","arguments":{"brief":"a jazz festival"}}}' \
  | vectorcraft-cli mcp --headless

Resource templates

vectorcraft://document (summary), vectorcraft://document/json (the whole model), and vectorcraft://commands (the live command catalog with enabled state) are the fixed resources. The templates read one thing at a time, which matters on a real document: the full model is thousands of lines an agent pays for again on every change.

uriTemplate	Reads
vectorcraft://object/{id}	One layer or object with its children, bounds and paint: document.node {id, summary: true}, the node as document.inspect lists it. A large container can be sliced with run_command document.node {id, summary: true, depth?, childLimit?} (a level that shows fewer children reports childCount)
vectorcraft://command/{id}	One command: label, menu path, shortcut, parameter description, enablement
vectorcraft://effect/{id}	One live effect with its parameters and defaults
vectorcraft://swatch/{name}	One swatch, colour, gradient or pattern swatch (percent-encode spaces)

An unknown URI is -32002; a template with no value, or a value that names nothing, is -32602.

The same reads work as query commands through run_command: document.node {id} returns the object's full model JSON (geometry, appearance, every attribute) and document.node {id, summary: true} the compact summary document.inspect gives for it (id, name, kind, bounds, paint labels, children), without summarizing the whole document. A summary read slices a large container: depth is how many child levels it includes (0: the node alone) and childLimit how many children each node shows, top of the stack first (both default to all). A level that shows fewer children than it has reports their total as childCount, so truncation is never silent; with nothing truncated the reply is the plain summary. Both must be non-negative integers and need summary: true (the full object JSON is never truncated); anything else is an error.

In the summary an object reports its paint as fill, stroke (labels) and strokeWidth. So does a compound path, whose paint is its own: its member paths report compoundMember: true instead of a paint they don't have. Every node (layers and groups too) reports opacity when it is below 1 and blend when it isn't Normal, as the label object.setProps {blend} takes ("Multiply", "Color Burn").

On a large document, drill instead of dumping: inspect_document {depth: 0} (or run_command document.inspect {depth?, childLimit?}, same rules) is the skeleton (artboards with top layers and counts), run_command document.find {name?, kind?, text?, limit?} locates nodes across the whole tree, and document.node reads the interesting ones sliced. document.inspect and the doc_inspect / inspect_document tools also scope the reply themselves: ids: [id…] lists just those nodes (any level, in the order given, an unknown id is an error, at most 10000), kind lists every node of one panel label (ignoring case; within the subtrees of ids when both are given) with its ancestor path, and limit / offset page the list. A scoped reply carries the nodes as nodes instead of layers (each the summary the full tree gives it, sliced by depth / childLimit) plus total, the entries before paging; artboards, selection and history stay whole. limit / offset alone page the top layers (layers and total). Without these options the reply is the whole layer tree as before. document.find needs at least one filter and all given filters must match; matching ignores case, name is a substring of the Layers panel name, text a substring of type content and kind the exact panel label (Layer, Group, Clip Group, Path, Compound Path, Type, Image, Rectangle, ...). It answers {matches: [{id, name, kind, path}], total} top of the stack first, where path is the ancestor chain (layer first) and limit caps matches (default 100; 0 only counts). document.json stays whole: it is the fidelity path, not the way to look around.

Completions

completion/complete answers for a prompt argument (ref/prompt) or a template variable (ref/resource), ranked by prefix and then by substring, capped at 100 values with total and hasMore. The values are read from the live catalogues, so they follow the app: formats suggests what the engine writes, preset the Image Trace presets, effect the effect ids, {id} the layer and object ids in the document, {name} its swatch names. An unknown reference comes back empty rather than as an error, since a completion is asked for mid-typing.

Logging

logging/setLevel accepts any RFC 5424 severity — debug, info, notice, warning, error, critical, alert, emergency — and an unknown one is -32602 with the previous level left in place. From then on the engine's own log records arrive as notifications/message with level, logger and data. A null level turns it back off. Nothing is sent before the client asks, and records queued at the old level are dropped rather than replayed at the new one.

Installing the logger that receives those records is vectorcraft-cli's job, not the library's: a crate that installed one would override whatever logger an embedder had already set up. Embedding vectorcraft-mcp directly means calling vectorcraft_mcp::logging::install() yourself if you want log records to reach your client.

Long calls: cancellation, progress and a busy engine

A call runs to the end before its reply, and the engine runs one call at a time, in the order the lines arrived. A thread of its own reads the stream meanwhile, so the server still hears the client while an export runs (#2165, #2171, #2172):

• Cancellation. notifications/cancelled {requestId} (spec) takes a request out of the queue, or stops the one that is running, and that request gets no reply. A running export, flatten, bake or render notices at its next checkpoint (between pages and artboards, between flattened groups, while drawing objects), returns, and writes nothing: the document is exactly as it was and the engine is free for the next call. An unknown, finished or malformed requestId is ignored, and so is a cancel of initialize. A call without checkpoints (every ordinary edit is over in milliseconds) runs to its end, but its reply is dropped all the same. Against a running app (--connect) the server stops waiting and drops the connection, but the app cannot be told to stop what it started.
• Progress. A request with _meta.progressToken gets notifications/progress (spec) before its reply: the PDF writer reports "Writing page 2 of 5" with progress 2 and total 5, raster export reports each artboard, and the PDF flattening stage reports a message without a total. progress only grows, updates come at most every 200 ms (the last page always), and none follows the reply. A client that does not read fast enough misses updates, never the reply.
• A busy engine. ping, tools/list, resources/list, resources/templates/list, prompts/list and prompts/get need no engine and are answered at once, whatever is running. Anything else that arrives while a call has already run for two seconds is not queued behind it: it is answered with error -32003 (data.busy names the running call, its request id and how long it has run) and was not run, so the client can retry later or cancel the running call. Requests that arrive sooner than that queue and are answered in order, as before, so a client that sends a few quick calls together sees no difference. At most 64 lines wait; past that the server stops reading until the engine catches up.

vectorcraft_mcp::Server::serve takes the input and the output on separate threads (Send), and Server::with_busy_after changes the two seconds. Calling handle_line directly is as before: one line in, one reply out, no threads.

Not implemented

• resources/subscribe / notifications/resources/updated. It would also have to be honest about whose edits it sees: comparing the undo history's length misses edits made in the app window, which is the main reason to run against a live document. document.inspect's revision is the right signal — it is monotonic and already bumped when a document is replaced — but noticing an edit still means someone asking.
• Reading the document during a long call. Requests that need the engine (doc_inspect, resources/read, command_list) are refused as busy while an export runs, not answered from a snapshot.
• sampling/createMessage and elicitation/create. Both are server→client requests. Supporting them means being able to suspend a tool call mid-flight, which is what the 2026-07-28 revision's input-requests mechanism exists for.
• Streamable HTTP / SSE. stdio only.

completion/complete covers the "what could I pass here" half of the same problem, which is why it is here and subscriptions are not.

Tools

Common command tools

These names match the conventions in FilmCraft #28. The older names below remain listed because existing workflows use them.

Tool	Arguments and result
command_list	Optional filter (or query), menu (a menu path substring, levels joined with >), enabled_only, brief (no params text); returns the command array
command_run	id, optional params; runs the command through the existing backend. An unknown key in params is an error that names it and the nearest key the command takes (invalid parameters for file.new: unknown parameter widht; did you mean width?), so a misspelt key no longer runs the command with its default (see Parameter keys)
command_batch	steps: [{id, params?}] (at most 256), optional stop_on_error (default true), atomic and label; returns completed, failed, results: [{ok, result|error}] and created (the ids the draw_path / draw_shape steps made, in order). A step's id is a command id, or draw_path / draw_shape with that tool's arguments as params (checked as tools/call checks them; the step's result is the tool's, so "$N.id" is the new object). Each edit has its own undo step, or with atomic: true the batch's edits are one undo step named label (else after its first edit; edit.undoGroup.begin/end underneath); failures set isError. A param string "$N.path" or "$last.path" takes a value from an earlier step's result (see run below). Kept step results total at most 8 MiB; past that the batch stops with the reason. This is the MCP tool, not the engine command command.batch (that one takes commands: [{command, params}] and is what the Actions panel plays); use command_run with command.batch when replaying a recorded action
doc_inspect	Optional depth, childLimit, ids, kind, limit, offset; same summary as inspect_document
render_preview	Optional max_side (1–4096, default 1024) and artboard (0-based, default 0); inline PNG of that artboard without editing it. Artboards too large to render within the allocation bound return a tool error
ui_inspect, ui_screenshot	Connected desktop state/window capture; tool errors in headless mode

Every tool declares read-only, destructive, idempotent and open-world hints. File-writing screenshot, save_file and export are conservatively marked as mutations. Unknown top-level argument keys return JSON-RPC -32602; command failures and caught tool panics return isError content, and the server continues serving. Command params are passed through: the registry currently describes them in prose, so MCP does not guess schemas or silently remove keys.

Coordinates are points in document space: y points down, the origin is the first artboard's top-left, and a new document is 612 × 792 (US Letter). New objects become the selection. Most commands act on the selection or on explicit ids. A command that leaves out an object selected inside a selected group also leaves it out when ids list it together with a group that holds it.

Paint values (fill, stroke) accept "#rrggbb", "none", [r,g,b] (0..1), {"c","m","y","k"}, {"gray"}, or a full paint.setFill params object ({"gradient": …}, {"swatch": "name"}). For draw_shape and draw_path, omitted paint fields inherit the current new-art appearance. Both also style the new object in the same call: dash (lengths, or null for solid), cap (butt|round|square) and join (miter|round|bevel) go to stroke.set with strokeWidth; opacity (0–100 %) and name to object.setProps; effects: [{effect, params?}] are applied in order as apply_effect does, to the whole object. They are checked before anything is drawn (effect ids against effect.list), so a bad one draws nothing. The reply adds name, opacity and effects (each effect.apply result) when given; paint lists the stroke options applied. A gradient given to several objects fits each one's own bounds; paint.setFill / paint.setStroke with span: "selection" place a linear or radial gradient once on the union of the targets' bounds instead, so one ramp runs across them. For draw_shape and draw_path, omitted paint fields inherit the current new-art appearance. For a filled shape or path without an outline, pass an explicit fill and stroke: "none". Lines start with no fill and a visible stroke. The Fill/Stroke proxy commands run through run_command: paint.invert and paint.complement recolour the active proxy keeping each colour's model, paint.lastColor / paint.lastGradient re-apply the last solid colour or gradient, and paint.recent returns the recent colours that every paint command (and the eyedropper) feeds. paint.proxies returns what the proxies show: the fill and stroke, which one is active, and whether the selected objects' fills or strokes differ (fillMixed / strokeMixed, drawn as a "?" proxy).

Tool	Arguments	Notes
list_commands	{filter?, query?, menu?, enabledOnly?, brief?}	The command catalogue: id, label, menu, shortcut, params doc (left out with brief), enablement.
run_command	{command, params?}	Runs any command. Use it for everything without a dedicated tool.
inspect_document	{depth?, childLimit?, ids?, kind?, limit?, offset?}	Artboards, layer tree (ids, kinds, bounds, paint), selection, history, tool. A type node's fill, stroke and strokeWidth are the paint its characters show (its first run's); fills and strokes of the type object itself come as objectFill, objectStroke and objectStrokeWidth. Paint comes as a label ("#ff0000", "None", "G1 40% (#ff9999)" for a swatch at a tint); paint.proxies gives it as an object.
inspect_ui	{}	UI state. Remote mode only.
select_tool	{tool}	selection, directSelection, pen, rectangle, ellipse, polygon, star, lineSegment, …
pointer_gesture	{events:[{kind,x,y,mods?}], tool?, mods?}	kind is one of down, drag, up, move, doubleclick. Events go through the same path as the mouse.
draw_path	{points | d, closed?, fill?, stroke?, strokeWidth?, dash?, cap?, join?, opacity?, name?, effects?}	points is [[x,y],…] or [{x,y,in?,out?,smooth?},…]. d is SVG path data. The options are below.
draw_shape	{shape, …geometry, fill?, stroke?, strokeWidth?, dash?, cap?, join?, opacity?, name?, effects?}	rectangle/ellipse: x,y,width,height (plus radius for corners). polygon: cx,cy,radius,sides. star: cx,cy,radius1,radius2,points. line: x1,y1,x2,y2.
set_paint	{fill?, stroke?, strokeWidth?, ids?}	Applies to the selection (or ids) and becomes the default for new art.
press_key	{key, mods?}	Remote: a real key event. Headless: runs the command or tool bound to that shortcut, or sends the key to the busy tool (digits too: 5 while dragging with the Perspective Selection tool; e while dragging a Selection or Group Selection marquee).
type_text	{text}	Remote only.
invoke_menu	{command, params?}	Invokes a menu item by command id. Includes UI commands such as view.* and window.* in remote mode.
open_panel	{panel}	Remote only. panel is a panel id (layers, swatches, colorGuide, …, as window.panel takes) or its display label ("Color Guide"), in any case.
screenshot	{path?, scale?, artboard?, window?}	Returns MCP image content (image/png, base64) plus a text block. Renders the artboard; window:true captures the app window (remote only). With path the backend (the headless session or the app) writes the PNG there too.
open_file	{path}	Opens any readable file as a new active document: .vectorcraft/.drawcraft, .vctemplate, .svg/.svgz, .pdf/.ai, .ait, .eps/.epsf, .ps, .dxf, .emf, .wmf, PNG, JPEG, GIF, WebP, TIFF, BMP, Targa (.tga/.vda/.icb/.vst), PSD/PSB (an image opens as a document of its pixel size; a Photoshop file as its merged image). Templates (.vctemplate, .ait) open as a new untitled document. PDF, .ai and SVG files saved with Preserve Editing reopen as the document they carry. The reply is document.open's: its warnings say what didn't come in as it was (an EPS whose PostScript can't be read opens as its preview, and the warning names the PostScript error, the operator and the procedure). run_command document.formats lists the formats.
save_file	{path?}	Runs document.save: the document's own file in its own format (native .vectorcraft unless it was opened from or saved as SVG, PDF or a restorable .ai; then warnings say what that format loses). A path's extension picks the format (.vectorcraft, .vctemplate, .pdf, .svg, .svgz, .ai: a PDF carrying the native document, which reopens editable).
export	{path?, format?, scale?, artboard?, range?, selection?, outlineText?, options?}	svg, svgz, pdf, eps, dxf, emf, wmf, png, jpg, webp, gif, png8 (an indexed .png), tiff, bmp, tga, psd (layered), txt (the document's text), vectorcraft or template (a native template). The tool's format enum and document.formats list them, generated from the engine's format table. When format is omitted, it comes from the path's extension. PDF writes one page per artboard: all of them, or artboard (0-based) / range ("1-3, 5", 1-based); the other formats write one artboard. options carries more format options (e.g. {"quality": 80} for JPEG). selection: true exports the selected objects cropped to their bounds (the reply adds their bounds, and reports format and the encoder's warnings as a whole-document export does); outlineText: true writes SVG text as paths. Template layers are left out, live effects are kept, and exporting vectorcraft never changes the document's path. Without path the bytes come back as dataBase64. Both backends run the same document.export call.
add_text	{text, x?, y?, width?, height?, fit?, fitMinPercent?, path?, mode?, pathEffect?, size?, font?, color?, justify?, vertical?}	Point type at (x, y) (justify: left, center or right starts, centres or ends each line on x); area type with width/height (fit: none, autoHeight, shrinkText, as text.areaOptions); or path + mode (area/onPath) to flow text in or along a path, with pathEffect (rainbow, skew, 3dRibbon, stairStep, gravity); justify (also auto, justifyAll) aligns the paragraphs of every kind, as text.setStyle does later. vertical: true uses the existing vertical writing mode for every kind (default false, horizontal). Returns {id}, with overflow: true when the text doesn't fit its frame or path.
export	{path?, format?, scale?, artboard?, range?, selection?, outlineText?, options?}	svg, svgz, pdf, ai (the PDF-compatible .ai that Save writes, which carries the document when every artboard is written in order), eps, dxf, emf, wmf, png, jpg, webp, gif, png8 (an indexed .png), tiff, bmp, tga, psd (layered), txt (the document's text), vectorcraft or template (a native template). The tool's format enum and document.formats list them, generated from the engine's format table. When format is omitted, it comes from the path's extension. PDF and .ai write one page per artboard: all of them, or artboard (0-based) / range ("1-3, 5", 1-based). SVG, SVGZ and EPS write one file per chosen artboard. Rasters, DXF, EMF and WMF write one artboard, or one file per chosen artboard with options: {"useArtboards": true}. Rasters also take options: {"rect": [x, y, width, height]} (page points, y down): that area of the page at the given ppi/scale, without an artboard and without changing the document; it is refused with artboard, range or useArtboards, with a non-positive or non-finite size, and for non-raster formats. options carries more format options (e.g. {"quality": 80} for JPEG). selection: true exports the selected objects cropped to their bounds (the reply adds their bounds, and reports format and the encoder's warnings as a whole-document export does); outlineText: true writes SVG text as paths. Template layers are left out, live effects are kept, and exporting vectorcraft never changes the document's path. Without path the bytes come back as dataBase64. Both backends run the same document.export call.
add_text	{text, x?, y?, width?, height?, fit?, fitMinPercent?, path?, mode?, pathEffect?, size?, font?, color?, justify?}	Point type at (x, y) (justify: left, center or right starts, centres or ends each line on x); area type with width/height (fit: none, autoHeight, shrinkText, as text.areaOptions); or path + mode (area/onPath) to flow text in or along a path, with pathEffect (rainbow, skew, 3dRibbon, stairStep, gravity); justify (also auto, justifyAll) aligns the paragraphs of every kind, as text.setStyle does later. Returns {id}, with overflow: true when the text doesn't fit its frame or path.
apply_effect	{effect?, params?, ids?}	Appends a live effect. Without effect, returns the effect catalogue with parameters and defaults.
pathfinder	{operation, ids?}	unite, minusFront, intersect, exclude, divide, trim, merge, crop, outline, minusBack. For a live version, apply the pathfinder.* effect to a group (effect.apply groups several loose selected objects first).
transform	{ids?, dx?, dy?, rotate?, scale?, scaleX?, scaleY?, reflect?, shear?, origin?, copy?}	Runs move, rotate, scale, reflect, shear in that order. With copy, the first step duplicates.
create_graph	{type?, x, y, width, height, csv? | series?, categories?, rows?}	The nine Illustrator graph types: column, stacked column, bar, stacked bar, line, area, scatter, pie and radar. An empty CSV cell or a null in rows is a blank value; a number in straight quotes is a label. Edit later with graph.setData / graph.setType via run_command. graph.column {design, columnType: vertical|uniform|repeating|sliding, each, fractions: chop|scale} draws the columns or bars of the series picked with Group Selection (or seriesIndexes, or all of them) with a graph design saved by graph.design; with no design it returns the series' current design and options.
text_wrap	{ids?, offset?, invert?, release?}	Area type below the objects (same layer) flows around them.
undo / redo	{}	Many calls that change nothing record no undo step, for example object.align with nothing to move, object.move by 0 without copy and object.showAll with nothing hidden. undo (edit.undo) after such a call undoes the step before it. inspect_document lists the undo steps as history. run_command with edit.clearHistory removes every undo and redo step of the active document and returns {cleared}, the number of steps removed; the document stays as it is, and the command records no undo step.

For example, create editable Japanese area type with the engine's vertical writing mode:

{"name":"add_text","arguments":{"text":"こんにちは","x":40,"y":40,"width":160,"height":240,"size":20,"vertical":true}}

vertical also applies to point type and to path + mode (area or onPath). Omit it or pass false for horizontal writing. This is the same option as text.create / text.createInPath; it does not change the engine's existing glyph orientation or CJK composition.

For SVG/SVGZ, export's top-level outlineText: true writes text as paths; DXF takes it too (its type is outlined anyway unless preserve: editability). PDF instead uses options: {"advanced": {"outlineText": true}}; its default is false (live text).

With mods.shift, a marquee dragged with the Selection tool toggles the objects it reaches (the selected ones leave the selection, the others join it, the rest stays), and one dragged with Direct or Group Selection toggles the anchors inside it (a path left with none leaves the selection, one with all of them is selected whole). The Lasso adds anchors with mods.shift and takes them away with mods.alt. They run select.toggle {ids} and select.anchorsMany {items, mode: "set"|"add"|"toggle"|"subtract"}:

{"name":"pointer_gesture","arguments":{"tool":"selection","mods":{"shift":true},"events":[{"kind":"down","x":180,"y":80},{"kind":"drag","x":300,"y":200},{"kind":"up","x":380,"y":200}]}}

press_key {key: "e"} while a Selection or Group Selection marquee is being dragged makes the marquee select only the objects that lie entirely inside it, measured by their geometric bounds, and e pressed again switches it back to the default mode. The Selection tool selects a group only when the whole group lies inside; Group Selection selects the objects inside groups, each as a whole object. Both run select.set {ids} (select.toggle {ids} with mods.shift). The marquee also selects the artboard guides that lie entirely inside it, and no canvas guide. Each marquee starts in the default mode. To press the key during a drag, end one pointer_gesture after the drag event and send the rest in another with the same tool, which keeps the drag:

{"name":"pointer_gesture","arguments":{"tool":"selection","events":[{"kind":"down","x":80,"y":80},{"kind":"drag","x":220,"y":170}]}}
{"name":"press_key","arguments":{"key":"e"}}
{"name":"pointer_gesture","arguments":{"tool":"selection","events":[{"kind":"drag","x":230,"y":170},{"kind":"up","x":230,"y":170}]}}

With nothing being dragged, e chooses the tool bound to E.

Appearance stacks: an object can carry several fills and strokes (appearance.addFill, appearance.addStroke), indexed in paint order (0 is painted first, the bottom row of the Appearance panel). Cmd+/ runs appearance.addFill and Cmd+Alt+/ runs appearance.addStroke (press_key {key: "/", mods: {cmd: true, alt: true}}). paint.setFill, paint.setStroke, stroke.set, stroke.setAdvanced, paint.editGradient, paint.setGradientGeom and transparency.set take item to edit one of them; run_command appearance.setActiveItem {"index": n} makes that row the target of later calls that omit item (as clicking the row in the Appearance panel does) until the selection changes; the proxy of its kind (paint.proxies) and the Gradient tool's annotator then show that row. inspect_document reports it as paint.appearanceItem. Live effects take the same item to apply to one fill or stroke instead of the whole object: run_command effect.apply {"effect": "path.offsetPath", "item": 0}, and effect.remove, effect.setParams (visible toggles one) and effect.duplicate address that item's effects; apply_effect uses the active item. effect.list reports each object's item effects under applied[].items.

effect.apply {"effect": "path.outlineObject"} (Effect › Path › Outline Object) takes no parameters. Type that carries it is drawn, exported and expanded as its glyph outlines and stays editable type, and the effects after it reshape those outlines. On paths, compound paths and images it does not change what is drawn.

Errors (unknown tool, bad arguments, a disabled or failing command) come back as a normal result with isError: true and a message. The model can read the message and retry.

Swatches go through run_command: swatch.list lists every swatch with its colour group, kind and colour; swatch.edit {name, newName?, color?, mode?, global?, spot?} edits one (fills, strokes and text linked to a global or spot swatch follow its colour and name, as one undo step); swatch.delete {names, unlink?} deletes swatches and colour groups in one undo step (art using a deleted global swatch keeps its colour, unlinked). swatch.new {color?, mode?, spot?, global?, group?} saves a colour (into a colour group with group), gradient or pattern; swatch.newGroup {fromArtwork: true, toGlobal?, includeTints?} makes a colour group of the selected art's colours (by default as global swatches the art links to). swatch.move {names, to?, group?} reorders swatches or moves them into or out of a colour group (as dragging them in the Swatches panel does); give colour group names instead to reorder the groups. Dropping a swatch on art in the app runs paint.setFill (or paint.setStroke, the active proxy) with ids: [the object under the pointer]. The Swatches panel menu's commands: swatch.addUsedColors {selection?, global?}, swatch.unused (a query: the names Select All Unused selects), swatch.merge {names} (the first is kept), swatch.ungroup {name} and swatch.sortByKind. Document colour mode: file.new {colorMode: "cmyk"} starts a CMYK document with CMYK default swatches, and RGB colors applied to it (paint.setFill/setStroke colors, gradient stops and freeform points, paint.editGradient stops, paint.freeform.addPoint/setPoint, paint.sampleColor, object.mesh.setPointColor/addLine, the text.create and text.createInPath color, the text.setStyle fill, text.setRangeStyle fills and strokes, a shapeBuilder.merge hex fill (its fill: {color, keepModel: true} form keeps RGB), the symbol.adjust stain color, swatch.new {color} and {colors}, swatch.newGroup, swatch.editGroup and swatch.edit) are stored as CMYK, converted through the color settings (edit.colorSettings: the CMYK working space and intent, as Document Color Mode converts them); Gray stays Gray, and keepModel: true keeps a color as given. A color that swatch.edit or swatch.editGroup is given and the swatch already has keeps its model, as does a stop color that paint.editGradient is given and the gradient already has; a stop that names a swatch takes the swatch's color as it is, and swatch.new or swatch.edit with mode converts a color into that mode only. RGB documents keep colors as given. Harmonies, Edit Colors blends, inversions and Recolor Artwork keep each colour's model (a blend between models takes the document's). swatch.new {colors: [...]} saves several colours as swatches in one undo step. color.harmony {color, rule, steps?, variation?, amount?} answers what the Color Guide panel shows: the harmony rule's colours, base first, and per colour its row of 2·steps+1 variations (shades, cool or muted on the left, tints, warm or vivid on the right). With limitTo (a swatch library id or name, or "document" for the document's swatches) every colour snaps to the library's nearest colour (ΔE 2000), as the panel's Limit to Library does (ui.colorGuideLimit); Recolor Artwork's limitTo takes the same values.

Color Themes (local only: no online service) keeps five-colour themes in the preferences. colorTheme.save {colors: [...]} saves up to five colours as they are, colorTheme.save {color, rule} saves the five-colour theme a harmony rule makes from a base colour (base first; name? defaults to "Theme N", replace: "<name>" overwrites that theme in place). colorTheme.list answers {themes: [{name, colors: [hex], keys: [colour keys], rule?}]}, colorTheme.delete {name} removes one and colorTheme.addToSwatches {name} adds one to the Swatches panel as a colour group (one undo step).

{"name":"run_command","arguments":{"command":"colorTheme.save","params":{"color":"#2266aa","rule":"splitComplementary","name":"Harbor"}}}
{"name":"run_command","arguments":{"command":"colorTheme.addToSwatches","params":{"name":"Harbor"}}}

SVG Options: export to SVG takes them in options, flat or as {"svg": {…}}: styling (presentation, style, entities, css), outlineText, images (embed, or link: embedded images are written next to the SVG, or returned as linked), objectIds (layerNames, minimal, unique), decimals (1–7), minify, responsive, useArtboards, range: "all" (one SVG per artboard, listed in files, each holding only the art over its artboard), preserveEditing (the SVG reopens as the full document), metadata and fewerTspans (one <tspan> per line of type). Unknown keys inside svg are rejected; run_command document.formats lists every option with its default. encoding is utf8, utf16 (big-endian after a byte order mark) or latin1 (ISO 8859-1, other characters as &#x…; references); document.serialize still answers text, plus dataBase64 (the file) when it isn't UTF-8, and such files open again. profile: "tiny12" writes a simplified SVG Tiny 1.2 (presentation attributes only; filters, masks, blend modes and embedded fonts left out, symbols as art, each with a warning). embedFonts: true embeds the fonts type uses as @font-face rules subset to the characters used; a font whose licence (OS/2 fsType) forbids subsetting is embedded whole, one that forbids embedding is left out with a warning. Type names its faces by numeric font-weight (600 for Semibold and for a Japanese face's W6, 300 for Light, bold for 700), embedded or not. svgz takes the same options and writes the SVG gzipped (.svgz files also open, place and paste). run_command document.save {path: "x.svg", svg: {…}} saves as SVG (or .svgz).

Symbols export as one <symbol> with a <use> per instance. An instance the def can't stand for is written as its own art: one stained by its fill; one scaled while the symbol has strokes (their weight doesn't scale), or rotated or scaled while it has effects, brushes or live objects; and every instance of a symbol with pattern paints or unlinked opacity masks (those stay on the page). hiddenLayers: true keeps hidden layers and objects, not displayed (display="none"); document.save and the CLI's convert keep them unless told otherwise, other exports leave them out, and they reopen hidden. An object's own data (object.setProps {data: {pivot: "100,180"}}) is written as data-* attributes (data-pivot="100,180"), and an SVG's data-* attributes come in as it. A preserveEditing SVG carries the native document (CDATA in <metadata>) and a hash of the markup around it: if another app changed the SVG since, document.open reads it as plain SVG and says so in warnings.

Saving beyond save_file goes through run_command: file.saveAs {path?, format?, options?, svg?} (the document takes on the new file and format; for format: "ait", or a .ait path without format, it fails with an error that names file.saveAsTemplate), file.saveCopy (the document keeps its path, title and modified state), file.saveAsTemplate {path?, format?} (it writes a .vctemplate, or, for format: "ait" or a .ait path, a PDF-compatible .ait written as a .ai file is, and with aiVersion: "7" it fails with an error; both open untitled; a path whose extension names another format has it replaced by the template's extension, and any other path that does not end in that extension has it added, so card.png is saved as card.vctemplate and card as card.vctemplate; the suggested name and the Templates folder come back without a path), file.newFromTemplate {path}, file.revert and file.formatOptions {format?} (a format's options with the values a save would use). Without a path the save commands return {dataBase64, name, folder?, warnings}. A document saved as SVG or PDF remembers its options, so the next save reuses them. A PDF opened with pages, cropTo or password is only part of the file, so Save asks for a new name instead of writing it back. file.documentColorMode {mode, convert?, intent?, grays?} switches the document colour mode, converting the colours of the art, symbols, pattern tiles and swatches through the colour settings (object.convertDocumentColorMode is an alias kept for older scripts). Gray colours stay Gray and print on the black plate only. RGB greys are colours like any other: to CMYK they separate through the profile into four-colour greys and a rich black, as Illustrator converts them. grays: "black" puts RGB greys with R = G = B on the black plate instead, K = their grey value (the inks Edit Colors › Convert to Grayscale before the switch would give); other colours convert as usual.

SVG import keeps what the canvas edits live. <pattern> becomes a pattern swatch. A <symbol> with <use> becomes a symbol (named after its data-name, else its id) with an instance per <use>; a <use> that shows it differently (paint inherited from the <use>, a viewport that cuts the art) or that the canvas would draw otherwise (a scaled symbol with strokes) stays art. feGaussianBlur, feDropShadow and the usual shadow chains (offset, blur, flood or colour matrix, composite, merge) become Gaussian Blur, Drop Shadow, Outer Glow, Inner Glow and Feather effects; other filters are listed in warnings. Undisplayed objects (display:none) come back hidden, nested clip paths intersect, and reflected or repeated gradients (spreadMethod) are expanded into stops. <image> files (relative links are found in the SVG's folder; only files on this computer or the SVG's network share are read, and pasted SVG reads only absolute paths) stay linked images (see Linked images); SVG files and data: SVGs become art; a file that can't be read becomes a placeholder in its box, named in warnings and missingLinks.

Native files: run_command document.save {compress: true} writes a gzip-compressed .vectorcraft (the useCompression preference makes that the default for saves; compressed files open like any other, told apart by their content). version: 2 or 1 writes a file older VectorCraft versions open (format name drawcraft, never compressed), and preview: true embeds a PNG of the first artboard, at most 256 px on its longer side. A save writes a temporary file and renames it over the old one, so a failed save never damages the file; only the images the document uses are written, and document keys written by a newer version are kept.

Resources

Fixed resources and the four templates (same as Resource templates under Protocol):

URI	Content
vectorcraft://document	document.inspect summary (JSON)
vectorcraft://document/json	The complete document model (JSON)
vectorcraft://commands	Live command catalog with enabled state
vectorcraft://object/{id}	One layer or object (document.node {id, summary: true})
vectorcraft://command/{id}	One command's metadata and enablement
vectorcraft://effect/{id}	One live effect with parameters and defaults
vectorcraft://swatch/{name}	One swatch (percent-encode spaces)

Examples

Draw, look, export:

{"name":"draw_shape","arguments":{"shape":"rectangle","x":72,"y":72,"width":200,"height":120,"radius":12,"fill":"#1e88e5","stroke":"none"}}
{"name":"draw_shape","arguments":{"shape":"star","cx":400,"cy":300,"radius1":90,"radius2":40,"points":5,"fill":"#ffc107","stroke":"#000000","strokeWidth":2}}
{"name":"draw_path","arguments":{"d":"M72 400 C 150 300 250 500 330 400","stroke":"#e53935","strokeWidth":4,"fill":"none"}}
{"name":"screenshot","arguments":{"scale":0.5}}
{"name":"export","arguments":{"path":"/tmp/art.svg"}}
{"name":"export","arguments":{"path":"/tmp/art.png","options":{"ppi":144,"background":"white","antiAlias":"type"}}}
{"name":"export","arguments":{"path":"/tmp/icon.png","range":"1-3","options":{"useArtboards":true}}}

Export As goes through options too: for PNG, JPEG and WebP useArtboards: true writes one file per artboard (all, or range), named <file>-<artboard>.<ext>, and returns {path, files} (a PDF keeps them as pages of one file; an SVG writes one file per artboard of a range); useArtboards: false exports the bounds of the visible art with what its effects paint, the effects of groups, type and images included. artboard.fitToArt measures the art the same way.

Raster exports take more options in options: ppi (72, 150, 300…; stored in the file), background (transparent, white, black or "#rrggbb"), antiAlias (none, art, type: text snapped to pixels) and, for PNG, interlaced.

Long-tail commands:

{"name":"list_commands","arguments":{"filter":"align"}}
{"name":"run_command","arguments":{"command":"select.all"}}
{"name":"run_command","arguments":{"command":"object.align","params":{"horizontal":"left"}}}
{"name":"run_command","arguments":{"command":"object.group"}}
{"name":"run_command","arguments":{"command":"document.exportPdf","params":{"path":"/tmp/art.pdf","range":"1, 3","compatibility":"1.5"}}}
{"name":"run_command","arguments":{"command":"document.pdfSettings","params":{"marks":{"trim":true},"includeDocument":true}}}

PDF files take the Save PDF dialog's options: preset, standard, compatibility, the General toggles and the compression, marks, bleed, output, advanced and security sections (list_commands with filter exportPdf documents every field). document.exportPdf and export (format pdf, the options in options) return warnings: options accepted but not applied yet, and features approximated or left out. An image whose pixel data can't be decoded (a PNG cut short) is left out of the pages, and warnings holds the note "an image couldn't be decoded and is missing from the picture" once, as raster exports give it; .ai saves and file.print do the same. A standard with a PDF version it doesn't allow is refused (PDF/A-2b at 1.3 or 2.0, PDF/X-4 above 1.6, PDF/X-1a and PDF/X-3 above 1.4); choosing a standard without giving compatibility sets its version (PDF 1.3 for PDF/X-1a and PDF/X-3). A security password encrypts the file (RC4 40-bit at PDF 1.3, RC4 128-bit at 1.4–1.5, AES-128 at 1.6, AES-256 at 1.7 and 2.0) with its permissions; either password opens it (document.open {password}), and a password with PDF/A or PDF/X is refused. Pattern fills and strokes are written as their tiles clipped to the area they paint (a stroke's outline, with its dashes, caps, profile, arrowheads and alignment), and freeform gradients as an image of their colour field at the document's raster effects resolution, clipped the same way. createLayers writes each layer and sublayer (template layers are left out) as a PDF layer, an optional content group named as the layer and listed under its parent layer's: hidden layers are off (their art is written, for the reader to show), non-printing ones have /PrintState /OFF and locked ones are locked. It needs PDF 1.5 or later (at 1.4 it warns), and the file reopens with those layers and sublayers. A .ai file writes them unless createLayers is false, so apps that read its PDF part find the layers, hidden ones included; hidden objects aren't in the PDF part. Images follow the compression settings of their kind (color, gray, or mono for black-and-white images): above abovePpi as placed they are resampled (downsample: average, subsample or bicubic) to ppi, and compressed with zip, jpeg (at quality; images with transparency stay lossless) or auto (JPEGs stay JPEG, the others lossless). none, jpeg2000, CCITT and runLength are written as ZIP, with a warning when an image needs them. CMYK images stay CMYK (DeviceCMYK, or ICC-based with the CMYK profile when colours are tagged, as in PDF/X-3 and PDF/X-4): a CMYK JPEG neither resampled nor recompressed is written unchanged, and the others' ink amounts are resampled and compressed again (CMYK JPEG or ZIP). output.conversion converts them like CMYK colours: destination to another CMYK profile or to RGB, preserveNumbers (and PDF/X-1a) keeps their numbers in a CMYK destination. document.pdfSettings lists the options that differ from the preset and the warnings without writing a file. thumbnails: true embeds each page drawn small (106 px on its long side, without the layers the page leaves out) as its /Thumb image. fastWebView: true writes a linearised file (the linearization dictionary first, then the first page with hint tables saying where every other page's objects are), which stays linearised when a password encrypts it. compatibility: "1.3" writes a PDF 1.3 file, which has no transparency: a copy of the document is flattened first with flattenerPreset (empty: High Resolution; high, medium, low or a saved preset, flattener.presets.list) and flattener options over it (as object.flattenTransparency takes them); images with see-through pixels count as transparency and rasterized areas are clipped to their regions. A file that would still have transparency fails the export. PDF/X-1a and PDF/X-3 files are PDF 1.3 files too. A page with a side longer than 14,400 pt is written with a UserUnit (PDF 1.6): its longer side divided by 14,400, rounded up to the next thousandth, with the page's sizes and coordinates in those units, so that no side is longer than 14,400 units (ISO 32000-1, Annex C.2). In PDF 1.3 to 1.5 files and PDF/X files, such a page is written in points.

An Illustrator EPS (version 9 on) or .ai file opens from the editing data it carries (after the EPS page; in the .ai file's PDF private data): layers and sublayers (name, order, visibility, lock, printing, preview, dimming, colour), groups as they were nested, compound paths, clipping groups, object names, hidden and locked objects, fills and strokes (spot colours as spot swatches), linear and radial gradients, opacity, blend modes, isolation and knockout, embedded images with their alpha channel, guides, and every artboard where it is. An object with several fills or strokes, effects or a brush comes in as its drawn look (a group named after it). Type is made from the file's text document, whole and editable: point type, area type (a story in several frames as threaded type) and type on a path, with its fonts, size, leading, tracking, scaling, baseline shift, fill and stroke (grey, RGB or CMYK), opacity, blend mode, alignment, indents and paragraph spacing. Where the text document can't say, type that shows is the page's, in its text object's place. A text object on a layer that shows and doesn't print keeps the type the text document makes for it, on that layer, even where the page doesn't draw it. A file whose editing data can't be decompressed or has symbols, pattern fills, placed files or anything else the reader doesn't read on a layer that shows, or whose layers look different from its page, opens as before: an EPS as its printed page, a .ai file from its PDF content (where plain groups open ungrouped), with a warning saying why. Layers look different when they draw an object the page doesn't (or miss one it draws), and, for an EPS whose page is the box of its art, when they print art outside it. A file whose editing data is for a page of another size opens the same way, and its warning gives both sizes in points, with the scale of the editing data's page to the file's page (1:10 for a tenth of its size) when their sides are in proportion. With textAs: "outlines" a .ai file whose type shows opens from its PDF content, which has the type's outlines. A PDF saved with Illustrator's editing data holds it in AIPDFPrivateData streams and opens from it as a .ai file does.

The document then has what the file's page doesn't print: hidden objects and layers, layers that don't print, guides and the art outside the artboards, as the app that wrote the file shows them. Exports and artboard.fitToArt leave out the hidden art, the guides and template layers; an EPS's artboard is the file's artboard, not the page's bounding box. A print pipeline that wants only what the page prints opens the file with editingData: false (default true): an EPS as its page, a .ai file from its PDF content (a .ai saved without PDF compatibility then can't be opened).

{"name":"run_command","arguments":{"command":"document.open","params":{"path":"/tmp/label.eps","editingData":false}}}

Opening a PDF (or .ai) imports every page as an artboard and layer; document.open takes pages ("2-3, 5", 1-based), cropTo (bounding (the art's bounds), art, crop (default), trim, bleed, media: the box each artboard gets; with trim the bleed box round it becomes the document's bleed, and a PDF-compatible .ai or a PDF saved with its editing data opens on its trim boxes for crop, its crop box taking the bleed in) and password for an encrypted file. document.pdfInfo reads a file without opening it: the page count, each page's size and boxes, needsPassword, and with thumbnail: n a PNG of page n. It reads a PostScript file (an EPS, a .ps file, a .ai in PostScript form) as one page the size of its bounding box (a US Letter page without one), with an empty boxes, and thumbnail: 1 shows what that page prints. A page whose UserUnit (PDF 1.6) is above 1 opens at its size in points, the file's numbers times the UserUnit (one above 75,000 counts as 75,000), and document.pdfInfo gives its size and boxes in points too. A .ai whose editing data has its artboards in the units of such a page opens from that data at the page's size. Imported colours keep their model: DeviceCMYK and CMYK ICC colours stay CMYK (a file painted mostly in CMYK opens as a CMYK document), DeviceGray is Gray, and Separation and DeviceN inks become spot swatches the art links to at its tint (gradient stops too). CMYK images (DeviceCMYK, or ICC-based with four components) keep their ink amounts: a CMYK JPEG as it is, any other as a CMYK TIFF (masked CMYK images, and ones with 1, 2 or 4 bits per sample, open in RGB with a warning). Placing a CMYK TIFF keeps it CMYK too. colorMode: "rgb" | "cmyk" opens any file in that mode instead, its colours converted as file.documentColorMode does (grays too):

{"name":"run_command","arguments":{"command":"document.pdfInfo","params":{"path":"/tmp/brochure.pdf","thumbnail":2,"cropTo":"trim"}}}
{"name":"run_command","arguments":{"command":"document.open","params":{"path":"/tmp/brochure.pdf","pages":"2-3","cropTo":"trim"}}}

PostScript files (.eps, .epsf, .ps, and .ai files saved in older formats or without PDF compatibility) open through the EPS reader (see EPS and PostScript import). An .ai saved without PDF compatibility (its PDF part is only a placeholder page) opens from its editing data alone: its type is made from the file's text document where it can be (point type, area type and type on a path), and what can't be is left out with a warning. Non-native art (the content of a placed PDF, which Illustrator shows but doesn't edit and keeps as a PDF inside the editing data) is drawn by the PDF importer, as a group named "Non-native art" fitted to the object's box; one whose PDF can't be read is left out with a warning. Without editing data it says it can't be opened. pages that name every page of such a file also open it from its editing data.

What a PDF holds comes in as editable art: soft masks become opacity masks (an alpha mask as a white copy of its art; the backdrop colour gives Clip, an inverting transfer function Invert), transparency groups keep isolation and knockout, tiling patterns become pattern swatches, patch and triangle mesh shadings become gradient meshes, and gradients keep their stop opacity and stop where the shading doesn't extend. Text becomes point type, one object per run of a line in the file's font (by name; fonts that aren't available are listed in warnings and show in the fallback font, and every export that draws that type — PDF, EPS, EMF/WMF, raster images, SVG with outlined or embedded fonts — says in its warnings that it wrote the fallback font) — textAs: "outlines" keeps the glyph outlines the file draws instead (its embedded fonts, installed or not). Lines wrapped in a frame (each wrapped line ending in a space, sharing a left edge, font and leading) come back as one area type object, left aligned or justified, where laying it out breaks the lines where the file does; a stroke written as each glyph's outline stroked is the type object's stroke; glyphs set apart by gaps wider than a space keep their places. Strokes stay live strokes (width, cap, join, miter limit, dash and paint), and an object written as a fill and then a stroke of the same outline is one path with both. Optional content groups (the layers of PDF and PDF-compatible .ai files) become layers with their name, visibility (the default configuration's, or a view state that is off), print state and lock, nested as sublayers the way the file's layer order nests them, stacked as the pages mark them (layers with no art come in empty); art that is off comes in as a hidden layer. Art outside them goes to a layer per page (except the opaque white page a .ai paints under its layers, which isn't art). layers: false gives one layer per page of only what shows:

{"name":"run_command","arguments":{"command":"document.open","params":{"path":"/tmp/map.pdf","textAs":"outlines","layers":false}}}

Preserve Editing (preserveEditing, on in the VectorCraft Default preset) embeds the native document in the PDF as an embedded file (vectorcraft-editing.vectorcraft; files named drawcraft-editing.drawcraft are read too) with a hash of the pages it was written with. open_file / document.open of such a PDF, .ai or .ait (no pages picked) restores the document exactly (restored: true); when another app changed the pages, or the data can't be read, the artwork is imported instead and the first warning says why. Choosing a standard turns it off (PDF/A refuses it). Save to a .ai path (document.save {path: "art.ai"} or {format: "ai"}) writes a PDF-compatible file that always carries the document, takes the PDF options and keeps its path when reopened, so Save writes .ai again; a .ait opens untitled. document.export / serialize to .ai (and the CLI's run --export x.ai and convert in.svg out.ai) write the same file without giving the document a path. options: {aiVersion: "7"} writes legacy .ai files of version 7 instead (the PostScript form, which keeps layers, groups, compound paths and clipping groups in apps that read it), one per artboard of range (default all; {name}-{artboard}.ai when several, listed in the result's files), with transparentImages: "flatten" (default) or "omit" for images with see-through pixels, opacity or a blending mode, hiddenObjects: "layer" (default: a hidden layer of each layer's hidden objects), "omit" or "visible", and outlineText: true to write type as outlines instead of text objects. When an image in art that is flattened or whose raster effects are expanded can't be decoded, the result does not show it, and warnings carries "an image couldn't be decoded and is missing from the picture" once. From the CLI: vectorcraft-cli convert in.svg out.ai --ai-version 7 [--hidden-objects layer|omit|visible] [--transparent-images flatten|omit] [--outline-text] [--range R]. document.export, serialize, exportSelection and exportForScreens refuse a .ait path or format: "ait" with an error that names file.saveAsTemplate, which writes it. The CLI's run --export x.ait and convert in.svg out.ait write it as file.saveAsTemplate does, with every artboard; convert refuses --artboard and --range for a .ait.

Drive a tool like a mouse:

{"name":"pointer_gesture","arguments":{"tool":"ellipse","events":[
  {"kind":"down","x":100,"y":100},{"kind":"drag","x":150,"y":140},{"kind":"up","x":200,"y":180}]}}

Edit a gradient with the Gradient tool's annotator (the bar runs along the vector; stop chips sit 10 px under it, midpoint diamonds 7 px above): a click on the bar adds a stop, dragging a chip moves it (drag it off the bar to delete it, hold Alt to copy it), dragging the end handle changes the vector. The selected stop (gradient.selectStop) takes Delete and ←/→:

{"name":"run_command","arguments":{"command":"paint.setFill","params":{"gradient":{"start":[100,150],"end":[200,150]}}}}
{"name":"pointer_gesture","arguments":{"tool":"gradient","events":[{"kind":"down","x":130,"y":150},{"kind":"up","x":130,"y":150}]}}
{"name":"pointer_gesture","arguments":{"events":[{"kind":"down","x":130,"y":160},{"kind":"drag","x":160,"y":160},{"kind":"up","x":160,"y":160}]}}
{"name":"press_key","arguments":{"key":"Delete"}}

A radial gradient's annotator also draws its extent: a dashed ellipse around the centre (the start, drawn as a ring) with a dot on it across the bar (drag it to change the aspect ratio); dragging the ellipse elsewhere rotates it. The dot inside the centre ring is the focal point, where the first stop sits: drag it for an off-centre radial, back onto the centre to centre it. paint.setGradientGeom sets the same things directly (aspect in %, focal in document coordinates or null; start/end may be left out), and they export as SVG fx/fy and PDF two-point radial shadings:

{"name":"run_command","arguments":{"command":"paint.setGradientGeom","params":{"aspect":60,"focal":[130,140]}}}

Freeform gradients: paint.editGradient {kind: "freeform"} places four or more colour points inside each selected object (coloured along the stops; mode: "points"|"lines" is the Draw toggle). paint.freeform.get lists the points (document coordinates), lines and the selected point; paint.freeform.addPoint {at, color?, opacity?, spread?, line?}, setPoint {index?, …}, deletePoint {index?}, addLine {points}, splitLine {line, segment, t?} and selectPoint {index|null} edit them (index defaults to the selected point; each edit is one undo step). The color of a point colored along the stops, of a point splitLine adds and of a point addPoint adds without color is a mix of the colors around it (the two stops, the line's two points, or all the points) in the color model those colors share, so CMYK stops and points give CMYK points. With the Gradient tool on a freeform gradient a click on a point selects it (drag to move it), a click on a line adds a point on it, a click elsewhere on the art adds a point (in Lines mode joined to the selected one), and Delete removes the selected point:

{"name":"run_command","arguments":{"command":"paint.editGradient","params":{"kind":"freeform","mode":"lines"}}}
{"name":"run_command","arguments":{"command":"paint.freeform.addPoint","params":{"at":[150,150],"color":"#ff3366","spread":20}}}
{"name":"pointer_gesture","arguments":{"tool":"gradient","events":[{"kind":"down","x":180,"y":170},{"kind":"up","x":180,"y":170}]}}

On the canvas the selected point also shows its spread as a dashed ring with a handle 16 px (or the spread, if larger) to its right: drag the ring or handle to change the spread. Dragging a point out of the object removes it, and double-clicking one opens its popover (dialog gradientStop, which on a freeform gradient edits the selected point: set color, opacity and spread and confirm). In Lines mode successive clicks on the art draw one smooth line through the points they add; a click on an existing point first continues the line from it, and press_key Escape (or a click off the art) ends it:

{"name":"pointer_gesture","arguments":{"tool":"gradient","events":[{"kind":"down","x":110,"y":190},{"kind":"up","x":110,"y":190},{"kind":"down","x":150,"y":170},{"kind":"up","x":150,"y":170}]}}
{"name":"press_key","arguments":{"key":"Escape"}}

Raw protocol (for debugging):

printf '%s\n' \
  '{"jsonrpc":"2.0","id":1,"method":"initialize","params":{"protocolVersion":"2025-06-18","capabilities":{},"clientInfo":{"name":"sh","version":"0"}}}' \
  '{"jsonrpc":"2.0","method":"notifications/initialized"}' \
  '{"jsonrpc":"2.0","id":2,"method":"tools/call","params":{"name":"draw_shape","arguments":{"shape":"ellipse","x":10,"y":10,"width":100,"height":80,"fill":"#ff0000"}}}' \
  '{"jsonrpc":"2.0","id":3,"method":"tools/call","params":{"name":"inspect_document","arguments":{}}}' \
  | target/release/vectorcraft-cli mcp --headless

Headless batch CLI

vectorcraft-cli commands                          # command catalogue (JSON)
vectorcraft-cli run --in in.svg \
  --cmd select.all --cmd paint.setFill --params '{"color":"#ff0000"}' \
  --export out.svg --export out.png --scale 2
vectorcraft-cli run --cmd file.new --params '{"width":800,"height":600}' \
  --cmd shape.star --params '{"cx":400,"cy":300,"radius1":200,"radius2":90}' \
  --export star.vectorcraft

run prints one JSON line per step (open, cmd, export) and exits non-zero on the first failure. --params applies to the --cmd just before it. run also accepts the host commands file.open, file.save, file.export, file.exportForScreens and tool.select. run --in, convert and info read every format document.open reads (vectorcraft-cli --help lists them).

What a headless session remembers

vectorcraft-cli run and the headless mcp server keep two things between runs, in folders next to the preferences (the folder the desktop app uses, so both see the same ones; VECTORCRAFT_CONFIG_DIR moves it):

• Saved PDF presets (pdf.preset.save, .delete, .import): one .vcpdfpresets file per preset in PDF Presets, written as the command runs. Presets saved earlier are there at the next start and work as preset in the PDF options.
• ICC profiles (color.loadProfile): a copy of the file goes into Color Profiles (the result says remembered: true), and every profile there is registered again at the next start, before the first command, so edit.colorSettings {"cmyk": "…"} or a preset's output intent can name it in a later run. The copy is what counts: the original may move or be deleted. color.removeProfile {name} forgets a profile and its copy (not while it is a working space).

Not remembered: Color Settings, the document, anything else. Nothing is read or written when VECTORCRAFT_NO_PREFS is set, and a server confined with --automation-read-root or --automation-write-root keeps nothing either (the folders are outside its roots). A file that is damaged, too large (a profile over 32 MB, a presets file over 1 MB) or not a valid preset or profile is skipped and logged. When the desktop app starts, the presets it kept in its preferences file before this folder existed are moved into it; after that the folder is the one place they live, so a preset deleted by vectorcraft-cli is gone in the app too.

Reproducible output

PNG, PDF, EPS, legacy .ai, packaged zips and saved .vectorcraft files record the time they were made in (PDF CreationDate and ModDate, the PNG Creation Time entry, a zip entry's date, a document's created and modified), so the same commands run twice gave different bytes though the pixels matched (#2880). Setting SOURCE_DATE_EPOCH to a non-negative whole number of seconds since the epoch records that instant instead of the clock, and the bytes repeat (https://reproducible-builds.org/specs/source-date-epoch/):

SOURCE_DATE_EPOCH=1791208980 vectorcraft-cli run --cmd file.new --params '{}' --export out.pdf

A command that takes its own date (file.new {created}, a save's modified) still wins over the variable. A blank, negative or non-numeric value is ignored and the clock decides, so an export is never refused over the environment. SVG output carries no time and was already reproducible.

Two things stay on the real clock and are not pinned by the variable: the recovery heartbeat, and the random seed PDF encryption draws on. A PDF exported with a password therefore still differs between runs whatever SOURCE_DATE_EPOCH says, because that seed supplies its AES initialisation vectors and its file /ID. Reproducible output means an unencrypted export.

Parameter keys

command_run, each step of command_batch, run --cmd … --params and the control channel's engine.execute check the keys of params against the command's params description (the params text command_list returns): the object it starts with, {x, width?: pt, name?|title?: …}, names the keys; ? marks an optional one. A key it doesn't name is refused before the command runs:

{"id": "file.new", "params": {"widht": 300}}
→ invalid parameters for `file.new`: unknown parameter `widht`; did you mean `width`?

Only the top level is checked: the keys of an object inside a value (artboardLayout: {…}) are the command's to read. The one exception is an effect's params: effect.apply (and appearance.addEffect) and effect.setParams check them against the effect's own entry in effect.list, its defaults and the keys its params text names. A key the effect doesn't take, or a value of another type than its default (a string for a number, a number for a switch), is refused the same way:

{"id": "effect.apply", "params": {"effect": "blur.gaussian", "params": {"blurr": 3}}}
→ invalid parameters for `effect.apply`: unknown parameter `blurr`; the command takes: radius

The color of Drop Shadow, Inner Glow, Outer Glow and Shift to Color must be one the effect can read: a hex color (#rrggbb, #rgb or #rrggbbaa, with or without the #), black, white, red, green or blue, [r, g, b] (three numbers, each read on a 0 to 1 scale), or a color object ({"model": "rgb", "r": 1, "g": 0, "b": 0}). effect.apply and effect.setParams refuse any other value themselves (where the parameter check runs first, a number or a boolean is refused as not a string), so the app's effect dialogs and Appearance panel, a step of command.batch and a replayed journal get the same error, and the document and its undo history stay as they were:

{"id": "effect.setParams", "params": {"index": 0, "params": {"color": "#ff00"}}}
→ invalid parameters for `effect.setParams`: `color` must be a #rrggbb color

A document saved with such a color opens as before, and the effect draws its default color (an array of three or more entries reads an entry that isn't a number as 0).

The channel of Levels and Curves must be rgb, red, green or blue (r, g, b and any letter case are read too). effect.apply and effect.setParams refuse any other value themselves in the same way, and a document saved with one opens as before and draws it as rgb:

{"id": "effect.setParams", "params": {"index": 0, "params": {"channel": "alpha"}}}
→ invalid parameters for `effect.setParams`: `channel` must be one of rgb, red, green, blue

A number outside the range the effect documents is still stored as given and clamped when the effect draws. A plug-in effect's parameters are resolved against its manifest instead, and effect.setParams as a step of command.batch isn't checked before the batch runs (the effect it edits may only exist once the steps before it ran); the command still refuses an effect color it can't read. A description with a … entry takes more keys than it lists: those of the command it names (…settings as file.new), or any key when it names none, as document.export and document.save do for the options of the format they write (one object can carry several formats' options). A command that takes its own reply back (document.setup, text.areaOptions) says so the same way. The app's own journal isn't checked: an action or journal recorded by an older version replays as it did, except for an effect color that the effect can't read.

These Object menu commands refuse a string or another value that is not a number with `scaleH` must be a number, not "50", naming the key, and change nothing: scaleH, scaleV, moveH, moveV and rotate in object.transformEach; ppi in object.rasterize; width, height, spacingX and spacingY in object.createObjectMosaic; offset, length, bleed and weight in object.createTrimMarks; and spacing in artboard.rearrange. A null is read as a key left out. object.cropImage refuses a rect that is not an array of four numbers, such as a string, with rect must be [x, y, width, height]; a null rect crops to the artboard, as a missing one does.

object.createObjectMosaic (columns, rows), object.transformEach (reference, seed) and artboard.rearrange (columns) take whole numbers, written 4 or 4.0. A fraction or a negative number is refused with `columns` must be a whole number. A string or another value that is not a number is refused with `columns` must be a number, not "4". Each message names the key that was given, and the command then changes nothing.

The blend, envelope and gradient mesh commands refuse a string or another value that is not a number with the same message, as in `steps` must be a number, not "8", and change nothing: steps, distance and value in object.blend.make and object.blend.options; bend, h, v, horizontalDistortion and verticalDistortion in object.envelope.makeWithWarp, resetWithWarp and options; rows, cols and columns in object.envelope.makeWithMesh, resetWithMesh and object.mesh.create; fidelity in object.envelope.options; highlight in object.mesh.create; and opacity in object.mesh.setPointColor. A null is read as a key left out. object.mesh.create refuses an at that is not two numbers with `at` must be [x, y].

A step can use an earlier step's result, in run as in the command_batch tool. A parameter string that is exactly $N or $N.path takes that value from the result of step N (counting from 0, in the order they ran; run --in's open isn't a step), and $last… from the step before. A path goes into objects by key and arrays by index: "$1.id", "$2.ids.0", "$last.ids[0]". A string starting with $$ stands for one $ ("$$1.99" is $1.99). A reference that names no value fails the step, so a command never runs with one left in it:

vectorcraft-cli run --cmd file.new --params '{"width":600,"height":200}'   --cmd text.create --params '{"x":10,"y":60,"text":"plain and bold words","size":30}'   --cmd text.setRangeStyle --params '{"id":"$1.id","start":10,"end":14,"style":"Bold"}'

Checking scripts written for the reference app

script.check {path? | data? | dataBase64?, name?, includes?} reads a script written for the reference app (.jsx, .jsxinc, .js, or a compiled .jsxbin) and reports what it uses. The check does not run any of the script, and it does not change the document or the preferences. The reply's kind is script, compiled (a .jsxbin file, or text that starts with @JSXBIN@, which the check doesn't read), module (a file with a top-level import or export, such as a React .jsx file), notText (a NUL byte, or more than one byte in ten a control character, in the first 8 KiB) or tooLarge (over 8 MiB). For a script, the reply holds these lists:

• needs lists what the script needs from the app that runs it: language (the script language's own objects and directives, such as $, alert, UnitValue and #target), dom (the reference app's document model), files, menuCommands, dialogs, and hostOnly (features only the reference app has: palettes, #targetengine, BridgeTalk, sockets, native code, XMP, the XML object and E4X, operator overloading, reflection, actions and selectTool).
• directives lists the #include, #includepath, #script, #strict, #target and #targetengine lines, and targets splits each #target line's application specifier into app, instance, version and locale. A directive is read where it starts a line, written #name or //@name.
• features lists the platform features the script uses by id (such as file.write, scriptui.dialog, menuCommand or e4x), each with a count and its lines.
• dom.members lists the document-model member chains as the script writes them (app.activeDocument.layers[].name, with [] for an index and () for a call); a variable assigned from a chain is replaced by that chain, and a name written with \u escapes is read as the name it spells. dom.types lists the type names the script compares with .typename or creates with new, apart from the language's, the platform's and the script's own functions.
• menuCommands.strings lists the string literals the script passes to executeMenuCommand (app.executeMenuCommand('…')), each with a count and its lines, and menuCommands.dynamic counts the calls whose argument isn't a string literal.
• outside lists what the script does outside the document: filesWritten (with the open modes used), filesOpened (files opened in other apps), network, nativeCode, otherApps (messages to other apps), environment (environment variables), unseenCode (eval, new Function and $.evalFile) and preferences (the reference app's preferences changed).
• problems lists what would stop the script loading as written (strings, comments, regular expressions, template literals and XML literals that aren't closed, brackets that don't match, syntax newer than ES3 and unknown directives) and notes on its text (bytes that aren't valid UTF-8, mixed line endings, and nesting deeper than the check follows: more than 4,096 open brackets, or XML literals and the code in their braces nested more than 4,096 levels deep).

Each entry of targets, directives, includes and problems has the line it is on (a problem with the whole text has none) and, when it is in an included file, that file's path as file. features, menuCommands.strings and outside count the script and its included files together and list up to 20 of the script's own lines.

With a path and includes: true, the check also reads the files the #include lines name, and the files their own #include lines name. When the script or one of these files assigns a member chain or a File object to a variable, the check treats the variable as holding it in every file. The check looks for a name in the folder of the file that holds the #include line, then in each folder that the last #includepath line before it names, taken from that same folder. It reads only files inside the script's folder: it doesn't follow an absolute name, a name, folder or link that leads out of the script's folder, a name or folder longer than 4,096 bytes, a file outside the automation read root, or the JSINCLUDE environment variable, and it reads each file once. An include's status is checked, missing, notFollowed, cycle (the file it names includes the file the line is in, directly or through other files), notRead (includes weren't read: no includes, no path, or the web app), limit (one of the limits below was reached) or unreadable (a compiled file, or not text), and its path is the file found, relative to the script's folder. The check reads at most 8 MiB of a file and 32 MiB in all, follows 64 included files to a depth of 16 and looks up at most 1,024 paths for them, keeps 1,000 entries in each list, follows at most 1,000 variables that hold member chains and 1,000 that hold File objects (truncated is true when an entry or a variable was left out), and cuts the text it echoes from the script to 120 characters.

File › Scripts › Check Script… picks a .jsx, .jsxinc or .jsxbin file and shows its report, and File › Open and drag-and-drop show the report for those files too; the web app checks the script without its included files. Over the control channel, ui.checkScript {path? | data? | dataBase64?, name?} does the same, and with a path it checks the included files too.

{"name":"run_command","arguments":{"command":"script.check","params":{"path":"/Users/me/Scripts/Tidy Layers.jsx","includes":true}}}

vectorcraft-cli run --cmd script.check --params '{"path": "Tidy Layers.jsx"}'

Transparency and opacity masks

Transparency and opacity-mask commands take ids (or id), so they need no selection. Opacity is a percentage (0..100) in every command (transparency.set, object.setProps, appearance.setItem). transparency.info returns the Transparency panel's values, with null where the objects differ. Making a mask from one object gives it an empty mask and enters mask editing: art drawn then becomes the mask, and the mask commands act on the masked object until transparency.stopEditingOpacityMask. Saves and exports never include the editing layer.

{"name":"run_command","arguments":{"command":"transparency.set","params":{"ids":[12],"opacity":40,"blend":"Multiply"}}}
{"name":"run_command","arguments":{"command":"transparency.makeOpacityMask","params":{"ids":[12,15],"invert":true}}}
{"name":"run_command","arguments":{"command":"transparency.setOpacityMask","params":{"id":12,"clip":false}}}
{"name":"run_command","arguments":{"command":"transparency.info","params":{"ids":[12,20]}}}

Knockout Group has three states: transparency.set {knockout: "on"|"off"|"neutral"} (true = on, false = neutral, the default). In a knockout group each child hides what it covers of the children below it; neutral groups pass the enclosing group's setting to their children, off groups never knock out. knockoutShape: true makes the object's opacity and opacity mask scale how much it knocks out. transparency.togglePageKnockoutGroup and transparency.togglePageIsolatedBlending ({value?}) treat the whole page as a knockout or isolated group; they are saved with the document and undoable, and transparency.info reports them. PDF and SVG write knockout groups as soft-masked groups with the same look (the PDF export reports a warning).

Groups are not isolated unless Isolate Blending is on: a blend mode inside a group with opacity, a blend mode, an opacity mask, knockout or a clip reaches the art below the group, on screen and in PDF (a knockout group's elements then composite against the art below it). SVG has no non-isolated groups, so there such groups stay isolated.

transparency.viewOpacityMask {on?, id?} is Alt-clicking the mask thumbnail: the canvas shows only the mask of id (default: the mask being edited, else the first selected masked object) as greyscale coverage (white = opaque) and the mask is edited; called again it shows the artwork, still editing. It is view state of the open document: the canvas (and ui.screenshot) shows it, exports never do, and it ends when mask editing ends.

view.transparencyGrid {on?} (View → Show Transparency Grid; default: toggle) is view state of each open document, like the mask view: the grid shows behind the artboards of the documents that turn it on, on the canvas and in ui.screenshot, never in exports. It returns {on}, isn't undoable and doesn't mark the document changed.

view.globalRulers {global?} (View → Rulers → Change to Global / Artboard Rulers, Cmd+Alt+R; default: toggle) is saved with the document and undoable. With artboard rulers (a new document's) the rulers count from the active artboard's top-left corner, or the zero point set for it; with global rulers, from one document zero point. Command coordinates stay absolute either way. document.rulerOrigin {artboard?} returns the zero point the rulers use with that artboard active; document.setRulerOrigin {x, y, artboard?} and document.resetRulerOrigin {artboard?} set or reset the document's (global rulers) or that artboard's (artboard rulers, kept relative to it so it follows the artboard).

{"name":"run_command","arguments":{"command":"view.globalRulers","params":{"global":true}}}

Clipping masks

object.clippingMask.make clips the selected objects by the topmost one, which may be a path, a compound path or a text object (it loses its paint). Compound holes, even-odd fills, glyph outlines and the union of a group's members clip alike on screen, in raster export and in SVG and PDF. object.clippingMask.release turns the clip group into a plain group and keeps the clipping path with whatever paint it has.

A clipping path can be painted after Make (paint.setFill, paint.setStroke, stroke.set with its ids): its fill paints behind the clipped art and its stroke over it, not clipped, on screen and in SVG and PDF.

{"name":"run_command","arguments":{"command":"select.set","params":{"ids":[12,15]}}}
{"name":"run_command","arguments":{"command":"object.clippingMask.make","params":{}}}

layer.clippingMask.toggle {id?} is the Layers panel's clipping mask button: the top object of the layer id (default: the one highlighted layer or group row, else the one selected group, else the current layer) becomes its clipping path (unpainted, moved to the bottom of the layer, so art added later is clipped too); called again it releases the mask. It returns {clip}, and the Layers panel underlines clipping-path names.

Object selection

select.set {ids} replaces the selection, select.add {ids} adds to it, and select.toggle {id} or select.toggle {ids} toggles each target. Every id must be a non-negative integer naming an existing object. A malformed or unknown id is an error naming that value; the entire selection is left unchanged. All three commands return {count, ids} for the resulting selection, in selection order. An empty array clears the selection for select.set and leaves it alone for Add and Toggle.

select.key {id} makes id the key object that Align aligns to and Distribute Spacing spaces from (object.align without to, object.distributeSpacing with spacing), and select.key {} cancels it. The id must name a selected top-level object. Layers, compound-path members and objects inside another selected object are not top-level objects. Any other id is an error naming that value, and the key is left unchanged. A null id is an error too. A click with the Selection tool on a selected top-level object, while two or more objects are selected or ruler guides are selected with it, runs select.key {id}, and a click on the key object again runs select.key {}. On a single selected object with no ruler guide selected, or on a group selected with only its own members, select.key sets no key. With a key object (or a key guide), object.align moves the other selected ruler guides too: a vertical guide goes to the key's left edge, centre or right edge, a horizontal one to its top, centre or bottom (#2143). Aligning to the selection or the artboard leaves guides where they are. select.key {guide} makes a selected ruler guide (its guide.list index) the key instead, while one or more objects are selected (#2143). object.align then lines the objects up on the guide and the guide stays put: a vertical guide sets where their left edges, centres or right edges go and vertical alignment leaves them where they are, a horizontal guide the other way round. Selected anchor points align to it the same way. object.distribute {spacing} and object.distributeSpacing {spacing} space the objects from it (see Align to Artboard). A click with the Selection tool on a guide selected with objects runs select.key {guide}, and a click on the key guide again runs select.key {}. The key guide goes when it leaves the selection or no object is left selected, and an object key replaces it (and the other way round). It returns {key}, the key object's id or null, with guide, the key guide's index, when a guide is the key. document.inspect (the inspect_document tool) reports the key object as key, null without one, and the key guide as keyGuide. The ui.pointer and pointer_gesture replies list it as key beside selection. The command palette lists select.key as Cancel Key Object and runs it without an id.

edit.clear {ids?} deletes the exact selected objects, including individual compound-path members; their unselected siblings remain in the compound. Supplying ids works without a selection and overrides selected anchors or guides. Every explicit id is validated before any deletion; a malformed or unknown id leaves the document unchanged. An empty ids array does nothing. Layers themselves are kept, and selecting an ancestor together with its descendants deletes that subtree once. Without ids, direct-selected anchors or selected ruler guides retain their usual Clear behavior. Cut still removes the objects it copied, including a whole compound when a member is selected.

object.setProps with visible: false or locked: true deselects every selected object that is hidden or locked after the call: the selected objects it hides or locks, selected objects inside them, and selected objects that were already hidden or locked. The key object is cleared when it is one of them, and edit.undo selects them again with the key object. When the call changes nothing, it records no undo step and still deselects them. Other object.setProps calls leave the selection unchanged.

layer.move, node.move, layer.merge, layer.flatten, layer.collectInNew, dataset.select, dataset.next and dataset.prev also deselect every selected object that is hidden or locked after the call, such as objects moved into a hidden or locked layer or group, or objects that a data set hides. The key object is cleared when it is one of them, and edit.undo selects them again with the key object.

These Select menu commands select only the objects that select.all selects and the objects inside them: every select.same.* and select.object.* command except select.object.slices, select.nextAbove, select.nextBelow and select.recall. They leave out hidden and locked objects with the objects inside them, the objects on hidden, locked or template layers and sublayers and, in isolation mode, everything outside the isolated group or layer. When the isolated group or layer, or a group or layer it is in, is hidden, locked or a template layer, they also leave out everything in it, and select.all, select.allOnArtboard and select.inverse select nothing. select.object.bristleBrushStrokes selects the objects whose topmost stroke has a bristle brush (type bristle in brush.list), whatever the brush is named; a brush of another type with "Bristle" in its name does not count. select.nextAbove and select.nextBelow select the nearest such object above or below the first selected object among the objects beside it in its group or layer, passing over hidden and locked objects and sublayers. When there is no such object, the selection stays as it is.

Saved selections

Select → Save Selection… keeps the selected objects under a name, in the document: select.save {name?} (default the first free "Selection N"; an existing name is replaced with the current selection; at most 25 per document, names up to 255 characters) → {name}, one undo step. select.savedList lists the names in menu order, and select.recall {name} selects those objects again, except the ones deleted since and the ones hidden, locked or outside the isolated group or layer (see Object selection) → {count}. Update Selection: select.update {} replaces the objects of the saved selection last recalled in the document with the selected objects → {name}, one undo step. The command is enabled while objects are selected and the document holds that saved selection, under its own name or a name select.editSaved gave it; Undo and Redo do not change which saved selection is recalled. Edit Selection… renames and deletes: select.editSaved {name, newName?, delete?}, or several at once as {edits: [{name, newName?, delete?}…]} in one undo step, where every name is the name before the edit and the names must stay unique. In the desktop app the saved selections are listed at the bottom of the Select menu (select.recall1 … select.recall25) and in the Layers panel menu's Save Selection submenu (select.recall {name}). A file's saved selections are checked when it opens: past 25, blank or repeated names, and ids the document doesn't have are dropped.

Ruler guides

Ruler guides are numbered in the order they were made. guide.add {vertical, pos, artboard?} makes one (the x of a vertical guide, the y of a horizontal one, in points) → {index}. With artboard (an artboard's index) it is an artboard guide: it runs across that artboard only, moves with it (artboard.move, a pure move in artboard.setProps, artboard.rearrange), scales with it (artboard.setProps {scaleArt}), is copied with it (artboard.duplicate, artboard.move {copy}) and is deleted with it (artboard.delete); without, a canvas guide runs across the whole canvas. guide.list → [{index, vertical, pos, selected, artboard?}…] (artboard only for artboard guides). guide.select {indexes: [index…], toggle?} selects guides on their own (the art is deselected; toggle adds or removes them) → {selected}. guide.move {index, pos} puts one guide somewhere, guide.move {dx?, dy?, copy?} moves the selected ones (vertical guides by dx, horizontal ones by dy; copy leaves them and selects the moved copies), and guide.remove {index?} deletes one guide or the selected ones → {count}; each is one undo step. With guides selected, edit.clear (Delete) deletes them and object.nudge (the arrow keys) nudges them. View › Guides › Lock Guides (view.guides.lock) deselects them and the paths made into guides, in every open document, and refuses these commands until unlocked. While it is on, a command that selects art, such as edit.undo, edit.redo, object.exitIsolation, object.unlockAll, object.showAll or edit.paste, leaves the guides out.

view.guides.release {index?, ids?} (View › Guides › Release Guides) turns guides into paths → {count}. With index it releases that ruler guide, and with ids it releases those guide paths (and the guide paths inside them). Otherwise it releases the selected guide paths and ruler guides, and with no guide selected it releases every guide path. A released ruler guide becomes a straight path with no fill or stroke, at the top of the guide's layer (or where new art goes, for a guide on no layer). An artboard guide becomes a line across its artboard, and a canvas guide becomes a line across the bounds of all the artboards. index and ids are refused while guides are locked. The release is one undo step, and the released paths are selected. With the Selection, Direct Selection or Group Selection tool, a double-click with Cmd/Ctrl and Shift held releases the ruler guide or guide path under the pointer (pointer_gesture with a doubleclick event and mods: {cmd: true, shift: true}), unless guides are hidden or locked.

The Selection, Direct Selection and Group Selection tools pick a guide within the selection tolerance, over the art (an anchor on it comes first with Direct Selection), and drag the selected guides: mods.alt copies them, mods.shift snaps the dragged guide to the ruler's ticks, and otherwise it snaps to whole pixels, the grid or, with Smart Guides, the edges, centres (so the sides' midpoints) and anchors of the art and of the artboards; with Smart Guides off, View › Snap to Point pulls it onto anchors. In the desktop app a guide dragged out of a ruler snaps the same way and is made where it is released over the canvas (with the Artboard tool, as an artboard guide of the active artboard), a guide dropped off the canvas onto its ruler is deleted, and hidden guides (View › Guides › Hide Guides) can't be picked. An artboard guide is picked, drawn and snapped to across its artboard only.

{"name":"run_command","arguments":{"command":"guide.add","params":{"vertical":true,"pos":100}}}
{"name":"run_command","arguments":{"command":"guide.add","params":{"vertical":false,"pos":200,"artboard":0}}}
{"name":"pointer_gesture","arguments":{"tool":"selection","events":[{"kind":"down","x":100,"y":50},{"kind":"drag","x":140,"y":50},{"kind":"up","x":140,"y":50}]}}
{"name":"run_command","arguments":{"command":"guide.list","params":{}}}

Symbols, brushes, variables and Image Trace

These families are in the live catalog (command_list / vectorcraft://commands); the prompt trace-and-style already walks Image Trace. Short agent entry points:

• Symbols: symbol.list (each row includes registration 0..8 and exportType: movieClip|graphic), symbol.new {name?, registration?, exportType?, nineSlice?, nineSliceGuides?, ids?}, symbol.options {name?, newName?, registration?, exportType?, nineSlice?, nineSliceGuides?} (nineSlice bool turns 9-slice guides on or off; nineSliceGuides is [left, top, right, bottom], fractions 0..1 of the symbol size; rows of symbol.list carry nineSlice: the guides or null), symbol.place {name?, x?, y?}, symbol.breakLink {ids?}, symbol.replace, symbol.edit, symbol.duplicate, symbol.selectInstances {name}. symbol.new and symbol.duplicate {name, newName?} trim the white space around the name they are given, as symbol.options trims newName; a blank one gives "New Symbol" or " copy".
• The Brushes panel menu's Select All Unused: brush.unused (a query) lists the brushes no stroke paints with, in library order, and the panel selects them. Strokes in the art (opacity masks included), symbols, graph designs, pattern tiles and graphic styles count, and so do the strokes in the art of a brush that is used. The art of a brush that is not used does not count, so a brush whose own art is the only art that paints with it is listed.
• Brushes: brush.list, brush.get {name}, brush.apply {name, ids?}, brush.new {type, name?, params?, ids?}, brush.setCurrent {name}, brush.options {name, params?}, brush.delete {name} | {names}, brush.duplicate {name, newName?} | {names}, brush.freehand, object.expandBrush {ids?}. brush.new, brush.duplicate and brush.options trim the white space around the name they are given. A blank name gives the type's default name ("New Calligraphic Brush" for a calligraphic brush), and a blank newName gives "<name> copy" in brush.duplicate and keeps the name in brush.options. A Hue Shift colorization given to brush.new or brush.options without a key takes the color that the art's fills and strokes use most (on a tie, the one found first), or black when they use no solid color. brush.options changes only the fields of colorization and of an art brush's scale that params gives, so {colorization: {method: "hueShift"}} keeps a Hue Shift brush's key.
• Variables: variable.list, variable.define {name, kind: text|visibility|image|graph}, variable.bind {variable, ids?}, variable.makeTextDynamic, variable.makeVisibilityDynamic, variable.makeImageDynamic, variable.makeGraphDynamic, variable.unbind, variable.rename, variable.delete, variable.highlight (image values are file paths, graph values CSV as graph.setData). variable.lock {locked?} (Lock Variables) locks or unlocks the variables; without locked it toggles, and the lock is saved with the document. While the variables are locked, variable.define, variable.delete, variable.rename and the four variable.make…Dynamic commands fail with "variables are locked"; variable.bind, variable.unbind and the dataset.* commands still run (except a dataset.import whose header would define a variable, below), and variable.list reports locked.
• Data sets: dataset.list, dataset.new {name, values?}, dataset.set {name, values?}, dataset.rename, dataset.delete, dataset.capture, dataset.update, dataset.select {name}, dataset.next / dataset.prev (each dataset.select / next / prev is one undo step over every bound object), dataset.import {path|data|dataBase64|csv} / dataset.export {path?} (CSV: name, then one column per variable). dataset.import also reads tab-separated files, and files whose first header cell isn't name: every header cell is then a variable, and the rows take the first free names Data Set 1, Data Set 2 and so on. An import error about a row's values or name gives the file line the row starts on. A header cell that names no variable defines one, unbound, in the same undo step as the rows. #Shown defines a visibility variable Shown, @Photo defines an image variable Photo, and a cell without a prefix defines a text variable. A prefixed cell whose name is already a variable of the prefix's kind names it; one whose name is a variable of another kind fails the import. A % cell that isn't a variable's exact name (graph data files) is refused. The reply's defined lists the variables the import defined. While the variables are locked, an import whose header would define a variable fails with "variables are locked".
• Batch: action.batch {commands, source?: folder|datasets, folder?, destination?: none|saveAndClose|folder, destinationFolder?, overrideOpen?, overrideSave?, overrideExport?, includeSubdirs?, stopOnError?} plays an action on every file in a folder or every data set (the Actions panel menu's Batch…). Override Save and Override Export end the file name of a step with format in that format, in lower case and without a leading dot, when it is one of the format's extensions (.svg and SVG give .svg, .JPEG gives .jpeg), and otherwise in the format's first extension (png8 gives .png, template gives .vctemplate). A step without format keeps the source file's extension (.vectorcraft over data sets).
• Image Trace: imageTrace.presets, imageTrace.make {id?, preset?, params?, view?} (or imageTrace.makeAndExpand), imageTrace.setView, imageTrace.expand, imageTrace.release.
• Live Paint: livePaint.make, livePaint.merge, livePaint.fill {group?, point}, livePaint.strokeEdge, livePaint.info {group}, livePaint.release, livePaint.expand, and livePaint.gapOptions {ids?, detect?, size?, color?, closeGaps?} (Object › Live Paint › Gap Options). With detect: true, paint stops at the gaps that run from a loose end of a path to the nearest path, up to size wide (small 3 pt, medium 6 pt, large 12 pt, or a number of points up to 72); closeGaps: true adds an unpainted path across each of them. The options are saved with the group, and the reply and livePaint.info list the gaps up to size wide that no path closes, as [x0, y0, x1, y1]. livePaint.merge keeps the bottom group's options when its Gap Detection is on, and otherwise turns Gap Detection on with small when another merged group has it on. When Gap Detection is off in every merged group, it stays off. view.livePaintGaps (View › Show Live Paint Gaps, or Hide Live Paint Gaps while it is on) turns the drawing of those gaps on or off. While it is on and edges show, the canvas draws the gaps of each selected Live Paint group in that group's Gap Preview Color. The reply is the new state (true when on). The Live Paint Selection tool's options (tool.setOption {tool: "livePaintSelection", values}; double-clicking the tool opens them, dialog livePaintSelectionOptions) are selectFills and selectStrokes (both on), color (the highlight color, #rrggbb or a layer color's name; #ff2a2a) and width (the highlight's width, 1 to 20 screen pixels; 4). With selectStrokes off, Alt-click selects the face under the pointer as a click does; with selectFills off, a click selects the edge nearest the pointer as Alt-click does; with both off, a click selects no face or edge.

dataset.select, dataset.next and dataset.prev reply {name, applied, skipped}. They add overflow: [id…] when area type or type on a path whose text they rewrote doesn't fit, as text.setText does (see Fitting area type). They skip an object that is locked or inside a locked layer or group, and count it in skipped. They rewrite hidden objects and objects inside a hidden layer or group, which stay hidden unless one of the data set's visibility values shows them. When such a layer is shown again, its type holds the data set's text. They apply the values in the order the variables are defined. Of two variables of the same kind on one object, the one defined last wins. dataset.list, dataset.capture and dataset.update read hidden objects and objects inside a hidden layer or group as well.

{"name":"run_command","arguments":{"command":"symbol.list","params":{}}}
{"name":"run_command","arguments":{"command":"brush.list","params":{}}}
{"name":"run_command","arguments":{"command":"variable.list","params":{}}}
{"name":"run_command","arguments":{"command":"dataset.list","params":{}}}
{"name":"run_command","arguments":{"command":"dataset.capture","params":{}}}
{"name":"run_command","arguments":{"command":"dataset.select","params":{"name":"Data Set 1"}}}
{"name":"run_command","arguments":{"command":"imageTrace.presets","params":{}}}
{"name":"run_command","arguments":{"command":"command.batch","params":{"label":"Offset","commands":[{"command":"object.move","params":{"dx":10,"dy":10,"copy":true}}]}}}

When no object is bound to a variable that the data set has a value for, for example after dataset.import defines new variables without binding them, dataset.select, dataset.next and dataset.prev apply nothing and add warning, which the app's status bar shows: "nothing in data set A was applied: no object is bound to a variable it has a value for". A data set with a value for a bound variable gives no warning even when every object bound to it is locked; skipped counts those objects.

dataset.update {all?} writes what the bound objects hold now into the active data set, for each variable that the data set holds. A variable whose objects report no value, because they are locked, inside a locked layer or group, or deleted, keeps the data set's value. With all: true it also adds every other variable that the objects report a value for. The reply gives values, the number of values it wrote, and left, the variables it did not write.

Graphic styles

A graphic style holds an appearance (fills, strokes, effects) plus opacity, blend mode, isolate and knockout. graphicStyle.new {name?, id?} captures an object (a group without its own fills or strokes lends its topmost object's, type its characters'). graphicStyle.apply {name, ids?} gives the objects the style and links them; add: true adds the style on top of the existing appearance instead. Linked objects stay linked while they keep the style's look: editing their appearance or transparency breaks the link, and graphicStyle.redefine {name?, id?} updates only the objects still linked. graphicStyle.list returns each style with the ids linked to it and the style of the first selected object; select.same.graphicStyle selects an object's fellow users. Styles keep placed gradients relative to the bounds of the object they were made from, and each object a style is applied to gets them at the same place relative to its own bounds.

{"name":"run_command","arguments":{"command":"graphicStyle.new","params":{"id":12,"name":"Glow"}}}
{"name":"run_command","arguments":{"command":"graphicStyle.apply","params":{"name":"Glow","ids":[15,20]}}}
{"name":"run_command","arguments":{"command":"graphicStyle.redefine","params":{"id":12}}}
{"name":"run_command","arguments":{"command":"graphicStyle.list","params":{}}}

graphicStyle.merge {names, name?} adds a style stacking the fills, strokes and effects of two or more styles (each on top of the ones before it) with the first one's transparency; graphicStyle.move {name, to} reorders the list (what dragging a style in the panel does). Override Character Color (graphicStyle.setOptions {overrideCharColor}, the preference overrideCharColor, on by default) makes a style applied to type replace its characters' fill and stroke with the style's fills and strokes; off, the characters keep their colour under them.

Editing appearance stacks

effect.move {from, to, fromItem?, toItem?, copy?} reorders an effect or moves it between the object's effects (null) and a fill's or stroke's (an item index), as dragging its row in the Appearance panel does; copy: true copies it (Alt-drag). appearance.duplicateItem {index, to?} / {indices} and appearance.removeItem {index | indices} act on several fills/strokes at once, and appearance.showAllHidden makes every hidden fill, stroke and effect visible again. In remote mode, effect.dialog {effect, index, item} opens an applied effect's dialog prefilled (ui.dialog.confirm runs effect.setParams); see the control protocol for the "already applied" question.

Overprint

Each fill and stroke (and each character's fill and stroke) can overprint: its inks print over the inks below instead of knocking them out. object.setOverprint {fill?, stroke?, item?, ids?} sets it (groups set their contents, type its characters too; item aims at one appearance item) and attributes.info {ids?} reads it back (null where the objects differ). Overprint Preview (view.overprintPreview) and Separations Preview show it; rendering approximates it by darkening (the stronger ink of each plate wins, so a colour overprinting itself stays that colour). Older files that listed Overprint Black objects in the document get the flags on load. PDF export writes them with advanced.overprint: "preserve" (the default): each overprinting fill and stroke sets a graphics state with /OP true /op true /OPM 1 (/OPM 0 in PDF/A). With discardWhiteOverprint (document.setup, on by default) white ones knock out instead; "discard" writes none.

{"name":"run_command","arguments":{"command":"object.setOverprint","params":{"ids":[12],"stroke":true}}}
{"name":"run_command","arguments":{"command":"attributes.info","params":{"ids":[12]}}}
{"name":"run_command","arguments":{"command":"view.overprintPreview","params":{"on":true}}}

Revolve (initial 3D effect)

effect.apply {effect: "threeD.revolve", params: {...}, ids?, item?} applies a live Revolve to a path or compound path. effect.dialog opens compact options with artboard rotation and light gizmos, effect.setParams edits it, and effect.expandAppearance bakes it into shaded vector faces. Native saves keep the editable source; SVG/PDF exports bake the same faces the canvas draws. Only one Revolve is supported per profile.

Parameters: angle (degrees, 0..360, default 360), offset (points, 0..100000, default 0), edge ("left" or "right"), rotationX/Y/Z (degrees, defaults 0/0/0), perspectiveAngle (degrees, 0..160, default 0; an earlier perspective, 0..100%, is read as the angle that shows the same view), segments (8..128, default 64), shade (default true), lightColor (opaque #rrggbb, default #ffffff), lightIntensity (0..100%, default 80), lightRotation (degrees, -180..180, default 125.26; 0 on the object's right, counterclockwise), lightHeight (degrees, 0..90, default 30; 90 straight in front), lightBehind (default false), lightSoftness (0..100%, default 0), ambient (Ambient Light, 0..200%, default 25; an earlier lightAzimuth and lightElevation are read as the same light), cap (boolean, default true), and expandVisibleOnly (boolean, default true). A vertical-line profile needs a positive offset to have a radius. cap closes each circular endpoint of an open profile with one curved, uniformly shaded vector path. Body faces share that curved rim. Partial sweeps use sectors and leave their angular cut sides open. Closed or coincident-end profiles are unchanged. Missing cap enables caps; set it to false for open endpoints. Shaded Revolve surfaces automatically interpolate curved-surface lighting with vector gradients while preserving profile corners and flat caps. There is no separate smoothing parameter. shade: false disables all lighting. Visibility of intersecting surfaces is approximate. lights is null by default, retaining the earlier single-light controls. Set an array of one to ten objects to replace that light: {color: "#rrggbb", intensity: 0..100, rotation: -180..180, height: 0..90, softness: 0..100, behind: bool}. Defaults per light are white, 80%, 125.26 degrees, 30 degrees, 0%, false. Omitted fields use defaults; unknown nested keys and malformed lists are refused before mutation. ambient is shared and counted once. Set lights: null to restore the earlier controls. The list is saved in the native file and survives preview, undo and expansion.

{"name":"run_command","arguments":{"command":"effect.setParams","params":{"index":0,"item":null,"params":{"lights":[{"color":"#ff8040","rotation":135,"height":35,"intensity":70},{"color":"#4080ff","rotation":-45,"height":20,"intensity":40}],"ambient":20}}}}

lightColor changes only the earlier directional light; ambient stays neutral. Invalid command colors are refused before editing, and invalid loaded colors fall back to white. Colored smooth faces use ordinary RGB vector gradients; source gradients keep their placement and opacity with vector blending/masks. Pattern tiles retain neutral shading. The dialog's Lighting tab manipulates the same lightRotation and lightHeight parameters; its guide visibility and tab choice are UI state, excluded from saved effect parameters. expandVisibleOnly is saved with the effect. It trims covered portions of opaque solid surfaces only when running effect.expandAppearance, preserving visible interiors through openings. Set it to false to expand all faces. Live preview and SVG/PDF baking keep the full surface. Transparent/unknown paint, non-normal blending, later geometry effects or bounded visibility failures preserve all faces. See Revolve for the scope and limitations.

{"name":"run_command","arguments":{"command":"effect.apply","params":{"effect":"threeD.revolve","item":null,"params":{"angle":360,"offset":20,"edge":"left","rotationX":0,"rotationY":0,"rotationZ":0}}}}
{"name":"run_command","arguments":{"command":"effect.setParams","params":{"index":0,"item":null,"params":{"angle":180,"lightRotation":30,"lightColor":"#ff8040","cap":true}}}}
{"name":"run_command","arguments":{"command":"effect.setParams","params":{"index":0,"item":null,"params":{"expandVisibleOnly":true}}}}
{"name":"run_command","arguments":{"command":"effect.expandAppearance","params":{}}}

Extrude & Bevel, Inflate and Rotate

effect.apply {effect: "threeD.extrude" | "threeD.inflate" | "threeD.rotate", params: {...}, ids?, item?} applies the other 3D types to a path or compound path, live like Revolve (the source stays editable; SVG/PDF exports and effect.expandAppearance use the same shaded faces). An object takes one 3D effect, Revolve included.

threeD.extrude: depth (points, 0..2000, default 50), twist (degrees, 0..360, 0), taper (1..100%, 100), cap (true solid, false hollow), bevel (none, classic, rounded, cove, step, ogee), bevelWidth and bevelHeight (points, default 4), bevelRepeat (1..32), bevelSpace (0..90%), bevelInside, bevelBothSides. threeD.inflate: depth (points, default 0), twist (degrees, 0..360, 0), taper (1..100%, 100), cap (true solid, false open behind), volume (0..100%, default 50), bothSides. All three take rotationX/Y/Z (degrees, -180..180, default the Off-Axis Front position -21.88/-23.4/9.06), perspectiveAngle and the lighting parameters as Revolve. See three-d.md for the Position presets, the scope and what differs from the public documentation.

Edit Colors and Recolor Artwork

edit.colors.invert, edit.colors.toCMYK, edit.colors.toGrayscale, edit.colors.toRGB, edit.colors.saturate, edit.colors.adjustBalance and recolor.apply recolour everything inside the selection (or ids) in one undo step: fills, strokes, text, gradient stops, gradient-mesh points, embedded images (a recoloured copy of the image; linked images are left alone) and the tiles of pattern fills and strokes (a recoloured copy saved as a new pattern swatch, e.g. "Dots 2"; the original pattern is untouched). includeImages: false / includePatterns: false leave images or patterns out. The Blend commands also grade gradient meshes, which keep their shading.

{"name":"run_command","arguments":{"command":"recolor.colors","params":{}}}
{"name":"run_command","arguments":{"command":"recolor.apply","params":{"map":{"#ff0000":"#0055ff"},"includeImages":false}}}
{"name":"run_command","arguments":{"command":"edit.colors.adjustBalance","params":{"mode":"cmyk","m":-20,"k":10}}}

Recolor Artwork keys colours by identity, not hex: recolor.colors lists each colour's key ("rgb 255 0 0", "cmyk 0 100 100 0", "gray 40", "lab 55 60 -40"; RGB in 0–255 levels, CMYK and Gray in percent), so CMYK, Lab and RGB colours that look alike stay apart, and every colour parameter also accepts a key. recolor.reduce {colors?, method?, preserve?, limitTo?} groups the selection's colours into rows of similar colours (k-means in Lab, weighted by use; tints of a global swatch share its row; White and Black are preserved by default) and returns a map of rows [{from: [keys], to: key}] to edit and pass to recolor.apply {map, method?, limitTo?, group?, groupColors?, rename?}. method says how a row's colours take its new colour: exact, preserveTints (tints of the row's darkest colour stay tints), scaleTints (every colour becomes a tint as light, relative to the darkest), tintsShades (lighter and darker than the row's average become tints and shades) or hueShift (the most saturated colour takes the new colour, the others turn by the same hue). limitTo snaps new colours to a swatch library's nearest colour; recolor.randomize {map, order?, saturationBrightness?, seed?} shuffles or varies new colours; a row with exclude: true keeps its colours. swatch.editGroup {group, colors, rename?} rewrites a colour group in one undo step (art linked to its global swatches follows), and recolor.apply with group recolours the art and rewrites the group together.

{"name":"run_command","arguments":{"command":"recolor.reduce","params":{"colors":2,"preserve":{"grays":true}}}}
{"name":"run_command","arguments":{"command":"recolor.apply","params":{"map":[{"from":["cmyk 0 100 100 0","cmyk 0 40 40 0"],"to":"cmyk 100 50 0 0"}],"method":"scaleTints"}}}
{"name":"run_command","arguments":{"command":"swatch.editGroup","params":{"group":"Brights","colors":["#ff0000","cmyk 0 0 100 0"]}}}

Group, layer and type appearance

Groups and layers carry fills, strokes and effects of their own, as in the reference app: their fills and strokes paint every member's geometry, and their effects apply to the members as one piece (one combined drop shadow; a Transform or Warp moves or bends the whole group). Their Contents row (type: Characters) is a slot in the stack: fills and strokes above it paint over the members (characters), those below under them; new items go above it (on an image, below it). appearance.moveItem {"from": "contents", "to": n} puts the row above the bottom n items, and an item move takes contents to say where the row ends up in the same undo step; document.node shows the slot as appearance.contents_index. The appearance.*, effect.* and graphicStyle.apply commands edit the selected objects themselves (a group's own stack; layers through ids) or, with target: "contents", the objects inside the groups and layers. appearance.targetContents selects a group's members (double-clicking the Contents row). SVG and PDF export bake a group's own fills, strokes and geometry effects into paths; its raster effects stay a filter on the whole group in SVG and become an image of the whole group in PDF. A geometry effect on one of type's own fills or strokes (effect.apply with item) reshapes the whole outline that the fill or stroke paints, and SVG, PDF, EPS and the other vector exports write such type as paths. On type that holds inline graphics, those effects are drawn and exported only when the type has a geometry effect or perspective of its own. An image's own fills and strokes draw only when they carry a visible Convert to Shape or Outline Object effect: Convert to Shape sizes its shape from the image's box, Outline Object gives the image's frame, the effects after it reshape that shape, and the effects before it are left out. Its Contents row stands for the image; appearance.addFill and appearance.addStroke put a new item right below that row, so it paints under the image. SVG, PDF, EPS and the other vector exports write those items as paths, and effect.expandAppearance turns them into objects in a group with the image. Use Preview Bounds measures an image by its frame.

{"name":"run_command","arguments":{"command":"appearance.addFill","params":{"ids":[7]}}}
{"name":"run_command","arguments":{"command":"appearance.moveItem","params":{"ids":[7],"from":"contents","to":1}}}
{"name":"run_command","arguments":{"command":"effect.apply","params":{"ids":[2],"effect":"stylize.dropShadow"}}}
{"name":"run_command","arguments":{"command":"appearance.addStroke","params":{"ids":[7],"target":"contents"}}}

Swatch libraries

Swatch libraries are read-only sets of swatches computed in code (Web Safe 216, Grays and Neutrals, Earth Tones, Skin Tone Ramps, Pastels, Brights, Metallic Gradients, Perceptual Scales, Harmony Sets). swatch.library.list lists them (id, name, category, count) and swatch.library.get {library} returns one's swatches and colour groups in the swatch.list shape (library is an id or a name). swatch.library.add {library, names?} copies swatches into the document as one undo step: a colour group's name brings the whole group, a swatch's name the swatch alone, no names the whole library; swatches the document already has (same name and paint) are reported under existing and not added again. apply: "fill"|"stroke" also applies the first one, in the same undo step (what clicking a swatch in the library panel does). swatch.resetDefaults {replace?} brings back the missing default swatches.

{"name":"run_command","arguments":{"command":"swatch.library.get","params":{"library":"earth-tones"}}}
{"name":"run_command","arguments":{"command":"swatch.library.add","params":{"library":"earth-tones","names":["Clay"]}}}

swatch.library.save {path?, format?: "vcswatches"|"gpl"|"ase"|"css", names?, name?, user?} writes the document's swatches as a library (.vcswatches keeps colour models, global, spot, gradients and colour groups; .gpl is 8-bit RGB; a swatch exchange .ase file keeps solid colors in their own model (RGB, CMYK, Lab or Gray) as global, spot or process colors, and color groups, writes a tint swatch as the color it shows and leaves gradients out; CSS writes custom properties); without path it returns {data}, or {dataBase64} for .ase, and user: true saves into the user library folder of the desktop app (listed as category user, User Defined). The reply's count is the number of swatches written: .gpl and .ase hold solid colors only, and CSS leaves freeform gradients out. swatch.library.load {path? | data? | dataBase64?, name?} loads a .vcswatches, .gpl, swatch exchange (.ase) or color book (.acb) file, or another document's swatches, as a library to add from. From an .ase file it reads RGB, CMYK, Lab and Gray colors as global, spot or process swatches and keeps their color groups. A color book (such as a Pantone book you own) gives its RGB, CMYK or Lab colors named with the book's prefix and suffix ("PANTONE 185 C"), as spot colors unless the book marks them as process colors; .acb files are read, never written. A file in the user library folder, or a file with the same extension and bytes as one there, loads as that User Defined library (category user). swatch.library.copyToUser {library} copies a loaded library into the user library folder of the desktop app (a library file as it is; a document's swatches or a library loaded from data or dataBase64 as .vcswatches) under a name no file there has, and lists it as User Defined from then on; until the app quits, commands given its loaded id use the copy → {library, name, count, path, copied} (copied: false when the folder already held the same file). Without a user library folder the command is disabled.

Graphic style libraries

Graphic style libraries are read-only sets of graphic styles generated in code (Shadows and Glows, Outlines and Rules, Hand-Drawn, Gradient Finishes, Shape Effects, Blends and Transparency). graphicStyle.libraries lists them (id, name, category: builtIn, user or loaded, count) and graphicStyle.library {library} returns one's styles in the graphicStyle.list shape (library is an id or a name). graphicStyle.addFromLibrary {library, name? | names?} copies styles into the document's Graphic Styles as one undo step (no names: the whole library); styles the document already has (same name and look) are reported under existing, a name another look has gets a number, and the patterns the styles paint with come along. apply: true also applies the first one to ids or the selection (add: true adds its appearance on top instead), in the same undo step: what clicking a style in the library panel does. Library styles keep placed gradients in unit-box space, so each object gets them at the same place relative to its bounds.

{"name":"run_command","arguments":{"command":"graphicStyle.library","params":{"library":"hand-drawn"}}}
{"name":"run_command","arguments":{"command":"graphicStyle.addFromLibrary","params":{"library":"hand-drawn","name":"Crosshatch","apply":true}}}

graphicStyle.saveLibrary {path?, names?, name?, user?} writes the document's styles as a .vcstyles library (JSON: the styles unlinked from swatches, with their opacity, blend mode, isolate and knockout, and the patterns they paint with); without path it returns {data}, and user: true saves into the user library folder of the desktop app (category user, User Defined). graphicStyle.loadLibrary {path? | data? | dataBase64?, name?} loads a .vcstyles file, or another document's graphic styles, as a library to add from; a file in the user library folder, or a file with the same extension and bytes as one there, loads as that User Defined library.

Library panels

Each swatch or graphic style library opens in a read-only panel of its own (#2603), as in Illustrator, and several can be open at once. window.swatchLibrary {library} and window.graphicStyleLibrary {library} (remote mode, invoke_menu) open one floating, or bring its panel forward when it is already open, and answer {open, panel, name, count}; library: null closes every panel of that kind. The panel id is library:<kind>:<library id> (library:swatches:pastels, library:graphicStyles:hand-drawn, library:swatches:loaded/C:/Palettes/Brand.ase), and the panel commands take it like any other panel's: window.panel.float {panel, onto?, x?, y?} groups or moves it, window.panel.dock {panel} docks it, window.panel.close {panel} closes it. window.swatchLibrary.other {path} and window.graphicStyleLibrary.other {path} load a file and open it the same way.

ui.libraryPanel {panel, library?, view?, showFind?, persistent?} sets one panel: library shows another library of its kind in that panel (what Previous Library and Next Library do; a library open in a panel of its own is brought forward instead), view is smallThumb, mediumThumb, largeThumb, smallList or largeList, showFind shows or hides the find field, and persistent: true is the panel menu's Persistent: the panel reopens when the app starts, with its view and find field, where the saved layout had it (or floating). A library loaded from a file is read again from its path; one whose file is gone is left out, and the status bar says which. A closed persistent panel can still be set, so persistent: false stops it reopening. The reply is {panel, kind, library, view, showFind, persistent, open}, and ui.inspect lists every panel's settings under ui.library_panels.

{"name":"invoke_menu","arguments":{"command":"window.swatchLibrary","params":{"library":"pastels"}}}
{"name":"invoke_menu","arguments":{"command":"ui.libraryPanel","params":{"panel":"library:swatches:pastels","view":"smallList","persistent":true}}}
{"name":"invoke_menu","arguments":{"command":"window.panel.close","params":{"panel":"library:swatches:pastels"}}}

Libraries

The Libraries panel's libraries hold graphics, colours and character and paragraph styles that any document can use. They are kept on this machine: each is a .vclibrary file (JSON) named after the library (CON-.vclibrary for a library named after a Windows device such as CON) in the Libraries folder next to the preferences (none in the web app and headless sessions, where they last for the session). library.list lists them (id, name, the count of each kind, current, folder); library.get {library?} returns one's items (graphics with id, name, size and a PNG thumbnail in base64; colours with name, hex and color; styles with name and attrs). library.create {name?}, library.rename {library?, name}, library.delete {library?} and library.setCurrent {library} manage them; library is an id or a name, and without it the current library (the one the panel shows) is meant.

library.add {library?, kind, ids?, name?, color?} adds from the selection: graphic (a copy of the objects with the images, symbols, patterns, swatches, styles and brushes they use, as Copy takes them), fillColor / strokeColor (the first selected object's solid colour, else the default one, or color), charStyle / paraStyle (the selected text's attributes). An item the library already has is reported as existing. library.use {library?, kind, item, center?, to?} uses one: a graphic is placed centred on center (pasted with its resources; a swatch name the document gives another colour merges into the document's), a colour paints the selection's fill or stroke, and a style is added to the document under its name without the white space around it (numbered when that name is taken by other attributes) and applied to the selected text, each as one undo step. library.removeItem {library?, kind, item} removes one.

Items can be put in user-named groups, which the panel shows as collapsible sections after the ungrouped items (its search field filters every item by name; library.get lists them all to filter). library.createGroup {library?, name?, items?: [{kind, item}]} makes one (a name another group has, ignoring case, gets a number) with those items moved into it; library.renameGroup {library?, group, name} renames one; library.deleteGroup {library?, group} deletes one, its items staying in the library ungrouped; and library.moveItem {library?, kind, item, group?, before? | after?} moves an item into a group (without group, out of its group) and, with before or after (another item of its kind), just before or after that item in the order its kind is listed and shown in. library.moveGroup {library?, group, to} moves a group to place to of the list (0: first; past the end: last). Dragging items and group headers in the panel runs these. An item is in one group at most; library.get lists the groups as groups: [{name, items: [{kind: "color"|"charStyle"|"paraStyle"|"graphic", item}]}] (a graphic by its id), and each kind's items in the order the panel shows them.

library.export {library?, path?} writes a library (its colours, styles, graphics with their art and thumbnails, and its groups) as a .vclibrary file, the JSON the library folder keeps (without path: {data}, the file's text). library.import {path? | data? | dataBase64?, name?} adds the library of such a file as a new library and makes it the current one; a name another library has gets a number (Brand (2)). Imported files are checked: at most 10,000 items of each kind and 1,000 groups, names clipped and made unique, thumbnails kept only when they are small PNGs, and groups naming items that aren't there dropped. Files written before groups load as they are.

{"name":"run_command","arguments":{"command":"library.add","params":{"kind":"graphic","name":"Logo"}}}
{"name":"run_command","arguments":{"command":"library.use","params":{"kind":"graphic","item":"Logo","center":[300,200]}}}
{"name":"run_command","arguments":{"command":"library.createGroup","params":{"name":"Logos","items":[{"kind":"graphic","item":"Logo"}]}}}
{"name":"run_command","arguments":{"command":"library.export","params":{"path":"/tmp/Brand.vclibrary"}}}

Strokes on type

stroke.set without an item gives type its characters' stroke: weight, cap, join, miter limit, align and the dash options go to every run (the object keeps no stroke of its own); an item still edits an object-level stroke added with appearance.addStroke, which takes every stroke option (dashes, profile, opacity, blend) like a path's and paints under or over the characters as its place relative to the Characters row says. To stroke some characters, use text.setRangeStyle {id, start, end, strokeOptions: {weight?, cap?, join?, miterLimit?, align?, dash?, dashOffset?, alignDashes?}}. align: "inside"|"outside" keeps the characters' stroke on that side of the glyph outlines, drawn at twice the weight and clipped, while the type stays editable (beyond Illustrator); SVG export writes the text live and that stroke as glyph outlines over it. inspect_document reports each object's stroke as strokeOptions (type: its first run's) in stroke.set terms. stroke.set on a group leaves the images and symbol instances in it alone.

Inline graphics in type

A document symbol can sit in a line of type like a character (beyond Illustrator; like InDesign's inline anchored objects), e.g. mana symbols in card rules text. text.insertInline {id?, at?, symbol?, scale?: 1, shift?: 0} inserts one at byte offset at (default: the Type tool's selection, which it replaces) of text id (default: the text being edited), showing symbol (default: the Symbols panel's current symbol) → {id, caret}; Type → Insert Inline Symbol does the same at the caret. The graphic is one character, U+FFFC, in a run of its own: text.getRange reports it as {text: "\uFFFC", style, inline: {symbol, scale, baseline_shift}}, plain text (and copied text) shows U+FFFC, and text.editRange with styled runs can insert one. Typing over it, Backspace and Delete remove it whole; neighbouring text never merges into it.

Layout: the art is scaled uniformly to scale × the run's font size tall, its left edge on the pen and its vertical centre on the middle of the cap height raised by shift points. Its advance is the scaled art width plus the run's tracking, so it breaks like a word (no break between it and punctuation stuck to it) and justifies like a glyph; a graphic taller than the font's ascent or descent opens up its line. Its art follows the symbol (Redefine Symbol resizes it); a missing or deleted symbol leaves an empty slot one em square that draws nothing. Its em box is its run's, whatever its scale: Character Alignment moves it with its run's text (it is one of the line's largest characters only when its run's size is), and Top-to-Top leading spaces its line by its run's em box (Roman leading by its run's leading). In right-to-left text it is a bidi neutral (U+FFFC): it takes the direction of the text around it and the line's reordering places it. The canvas draws the art inside the type's transparency (its opacity and blend mode apply). SVG keeps the text live (the text after a graphic is positioned after it) and writes each graphic as a <use> of the symbol's <symbol> def (or a copy of its art); type on a path with graphics is written as outlines. PDF draws the art as vector paths where the layout puts it.

Flatten Transparency

object.flattenTransparency {ids?, preset?, …options} turns transparent art into opaque art that looks the same, in one undo step. The targets split into groups of overlapping objects; groups without transparency stay as they are. Each other group becomes one group of atomic regions: paths filled with the flat colour the art showed there (over white; over nothing, keeping alpha, with preserveAlpha: true), plus one image where gradients, patterns, images, opacity masks or raster effects reach (clipped to those regions with clipComplexRegions, a rectangle without). preset is high, medium (the default) or low; option keys (balance 0–100, lineArtPpi, gradientPpi, textToOutlines, strokesToOutlines, clipComplexRegions, antiAlias, preserveAlpha, preserveOverprints) override it, at the top level or in options. balance: 0 rasterizes everything; lower balances rasterize groups that split into many regions. The result reports {ids, vector, rasterized, options}. When an image in the art can't be decoded, the image that replaces the art does not show it, and the result also carries warning: "an image couldn't be decoded and is missing from the picture", the note raster exports give.

{"name":"run_command","arguments":{"command":"object.flattenTransparency","params":{"ids":[12,15],"preset":"high"}}}
{"name":"run_command","arguments":{"command":"object.flattenTransparency","params":{"balance":0,"lineArtPpi":150}}}

Pattern editing display

In pattern editing mode the tile edge (pattern.options {showTileEdge}) and the swatch bounds (pattern.options {showSwatchBounds}: the part of the tiling the swatch repeats, dashed) are drawn in the preference patternTileEdgeColor.

object.pattern.make {name?}, object.pattern.saveCopy {name?} and pattern.options {newName?} trim the white space around the name they are given. The first two number a name that a pattern, a swatch or a color group has ("Dots 2"), and pattern.options refuses a name that another pattern, a swatch or a color group has. It takes a name that only the pattern's own swatches have (a name Swatch Options gave the pattern's swatch, for example) when none of them still has the old name. A rename gives the new name to the pattern's swatches that had the old one; any other swatch that has the old name keeps it. A blank name gives "New Pattern N" in object.pattern.make and "<pattern> copy" in object.pattern.saveCopy, and pattern.options keeps the name.

{"name":"run_command","arguments":{"command":"pattern.options","params":{"showSwatchBounds":true}}}
{"name":"run_command","arguments":{"command":"prefs.set","params":{"key":"patternTileEdgeColor","value":"#ff4f4f"}}}

The Eyedropper: appearance.copyFrom {source} copies what the Eyedropper Options pick up and apply from source to the selection (ids); reverse: true copies the selection's attributes onto source (Alt-click) and append: true adds the source's fills and strokes on top of each target's stack (Shift+Alt-click). The options live in the preferences: eyedropper.setOptions {sampleSize?, pickUp?, apply?} (or prefs.get/prefs.set with key eyedropper) reads and sets them; a tree names flags under appearance (transparency, fill {color, transparency, overprint}, stroke {color, transparency, overprint, weight, cap, join, miter, dash}), character and paragraph, and a bool for a branch sets all of it. paint.sampleColor {color} puts a sampled color into the active proxy (in a CMYK document an RGB color is stored as CMYK, as paint.setFill stores it). Clicking an image samples its pixels: a CMYK image (a four-component JPEG, or an 8- or 16-bit chunky CMYK TIFF) gives its ink amounts as they are, and an RGB image's sRGB pixel is taken into the working RGB space.

{"name":"run_command","arguments":{"command":"eyedropper.setOptions","params":{"pickUp":{"appearance":{"stroke":{"weight":false}}}}}}
{"name":"run_command","arguments":{"command":"appearance.copyFrom","params":{"source":12,"ids":[7,8]}}}

The Layers panel: rows, layers and sublayers

Every layer, sublayer, group and object is a row of the Layers panel. Sublayers are layers inside layers: they are not objects, so clicking or marquee-selecting art in a sublayer, Select All and the other Select commands take the art itself, and each sublayer has its own colour for the selection highlight. document.inspect lists the tree (layers report color, template, printable, preview, dimImages and clip), the currentLayer (where new art goes) and layerRows, the rows highlighted in the panel.

• Clicking rows (not undo steps): layer.setCurrent {id} is a plain click on any row: it alone is highlighted and its layer (the row itself when it is a layer or sublayer) becomes current. layer.highlight {ids, mode?: set|add|toggle} is Shift-click (a range) and Ctrl/Cmd-click (toggle). Selecting art makes its layer current. Without ids, the panel commands below act on the highlighted rows, else on the current layer.
• Selection column: layer.selectAll {id, add?} selects a row's visible, unlocked art (a layer's sublayers' too); add (Shift) adds it, or removes it when it's all selected.
• Eye and lock: layer.setProps {ids?, visible?, locked?, name?} for any rows, plus Layer Options for layers: template (also locks the layer and dims its images to 50% unless given), printable, preview (off: the layer draws and is clicked in outline, Ctrl/Cmd-click its eye), dimImages (0–100, or false) and color (an index 0–26, #rrggbb or a preset name). Several rows change in one undo step (a drag down the eye or lock column).
• Dragging rows: layer.move {ids, target, place?: above|below|inside, copy?} moves rows beside a row or into a layer or group (on top of its contents), keeping their stacking order; copy (Alt-drag) moves copies. Layers go only in layers or at the top level and objects in layers and groups (an object placed beside a top-level layer goes inside it); a row never goes into itself or a row inside it, and a locked layer or group takes nothing. Dragging the selected-art square is layer.move with the selected objects. node.move {id, parent?, index} is the low-level form with the same nesting rules; unlike layer.move, it also moves a row into a locked layer or group.
• Buttons: layer.new {name?, top?, …options} (above the current layer at its level; top with Ctrl/Cmd), layer.newSublayer {parent?, name?, …options} (on top of the parent's contents), layer.delete {ids?} (rows with what they hold; the last layer stays), layer.locate {id?} (highlights the selected object's row; the panel opens the rows around it) and layer.clippingMask.toggle.
• Panel menu: layer.duplicate {ids?}, layer.merge {ids?} (into the layer highlighted last, keeping the stacking order), layer.flatten {id?} (every other visible layer's art into one layer; hidden layers are deleted, templates stay), layer.collectInNew {ids?}, layer.releaseToLayers {id?, build?} and layer.releaseToLayersBuild (each object of a layer or group in a sublayer of its own; Build adds up copies), layer.reverse {ids?}, layer.template {ids?, on?}, layer.hideOthers / layer.showAll, layer.outlineOthers / layer.previewAll (Ctrl/Cmd+Alt-clicking an eye runs Outline Others for the row's layer, or Preview All Layers when the others are in Outline already), layer.lockOthers / layer.unlockAll (Alt-clicking an eye or a lock runs Hide or Lock Others for the row's layer, or Show or Unlock All when the others already are), layer.pasteRemembersLayers, object.isolate {id} (a group, a layer or a single object; not type) / object.exitIsolation.
• Dialogs and panel state (control channel and the app's MCP): ui.layerOptions {ids?} and ui.newLayer {sublayer?} open Layer Options (dialog layerOptions: name, color, template, locked, visible, printable, preview, dimImages, dimPercent; OK is one undo step), ui.layersPanelOptions opens Panel Options (layersPanelOptions: layersOnly, locateSelected, rowSize small|medium|large|other, otherSize, thumbLayers, thumbGroups, thumbObjects), and ui.layersExpand {ids?, open?} opens or closes rows as their triangles do (Alt-click: everything inside). With locateSelected on (Locate Selected Objects, off by default), when the last selected object changes on the canvas or by a command, the panel opens the layers and groups around it and scrolls to its row, as after layer.locate, but does not highlight the row. When Show Layers Only, a search or the filter leaves the object without a row, the panel opens no rows for it and scrolls as with the option off. A change made with the panel's own rows, buttons or menu is not located, and a change made during a drag is located when the drag ends. A document opens with only its top-level layers open; the open rows are saved in the native file and reopen that way.
• Search and filter (control channel and the app's MCP): the Search All field keeps the rows whose names hold the text typed, and ui.layersFilter {objects?, layers?} sets the choices of the Filter button beside it. objects (By Object) takes paths (paths and compound paths), shapes (live rectangles, ellipses, polygons and lines), text, images, effects (rows with an effect on themselves or on a fill or stroke, shown or hidden), groups, linkedFiles (linked images and placed documents) and guides (paths made into guides, and ruler guides); layers (By Layer) takes locked and hidden (the row's own lock and eye). A row stays when it is one of those kinds and in one of those states (an empty list takes every row) and its name holds the search text, or when a row inside it stays. A key left out keeps its list. The filter lasts the session and does not change the document, its open rows or the selection. ui.layersFilter answers {objects, layers}.

{"name":"run_command","arguments":{"command":"layer.newSublayer","params":{"name":"Shadows"}}}
{"name":"run_command","arguments":{"command":"layer.move","params":{"ids":[12,15],"target":7,"place":"inside"}}}
{"name":"run_command","arguments":{"command":"layer.setProps","params":{"ids":[7],"preview":false,"color":"Orange"}}}
{"name":"run_command","arguments":{"command":"layer.highlight","params":{"ids":[3,7]}}}
{"name":"run_command","arguments":{"command":"layer.merge","params":{}}}

Targeting layers and moving appearances

layer.target {id} is the Layers panel's target circle: a layer gets its visible, unlocked art (its sublayers' too) selected and is itself the target, so appearance.*, effect.*, transparency.* and the opacity-mask commands without ids act on the layer (its opacity, its own fills and effects, an opacity mask on the whole layer); a group or object is simply selected. document.inspect reports it as target, and any other selection change ends it. appearance.transfer {source, target, copy?} is dragging a target circle onto another: the target gets the source's fills, strokes, effects and transparency, and the source is cleared unless copy (Alt-drag). Dropping a circle on the panel's trash is appearance.clear {ids: [id]}. Masked objects' names have a dashed underline; while an opacity mask is edited the panel lists only an <Opacity Mask> entry and the document tab says (<Opacity Mask>/Opacity Mask).

{"name":"run_command","arguments":{"command":"layer.target","params":{"id":2}}}
{"name":"run_command","arguments":{"command":"transparency.set","params":{"opacity":50}}}
{"name":"run_command","arguments":{"command":"appearance.transfer","params":{"source":12,"target":2,"copy":true}}}

Expand Appearance

effect.expandAppearance {ids?, target?} (Object → Expand Appearance, enabled when a selected or targeted object's appearance isn't basic) turns appearances into plain objects in one undo step: each visible fill becomes a copy of the path painted by that fill alone and each stroke its outline filled with its paint (a brushed stroke: its brush art), grouped in paint order under the object's id; each piece takes its fill's or stroke's opacity and blend mode, and the group keeps the object's transparency and opacity mask. Geometry effects are baked; raster effects become an embedded image at the document's raster effects resolution (shadows and outer glows under the art; a blur, feather or inner glow replaces the object with the image). Type with a geometry effect (Outline Object included) or with fills and strokes of its own is outlined; type that has no fills or strokes of its own and whose only effects are shadows or outer glows stays type, with the image under it. A group's or layer's own fills and strokes become objects among its members, and its members are expanded too. The result is {ids}; when an image in the art can't be decoded, the images made from raster effects do not show it, and the result also carries the warning that Flatten Transparency gives.

{"name":"run_command","arguments":{"command":"effect.expandAppearance","params":{"ids":[12]}}}

Tints of global and spot colours

A colour linked to a global or spot swatch has a tint (the reference app's T slider): paint.setFill {swatch, tint?: 0..100} (and paint.setStroke) applies the swatch at that percentage, linked, so swatch.edit recolours it at its own tint. paint.proxies shows the paint as {type: "solid", color, swatch, tint} (tint 0..1, left out at 100 %); document.inspect reports it as a label, the swatch name, the tint below 100 % and the colour it gives ("G1 40% (#ff9999)", "G1 (#ff0000)"). swatch.new {tint?} with the current fill (or swatch) saves a tint swatch, "Name 40%": swatch.list reports it with tintOf and tint, applying it links to its base at that tint, and it follows edits of its base. A spot tint prints that percentage of its plate (Separations Preview, PDF Separation value). edit.colors.adjustBalance {tint: -100..100} (Global mode) shifts the tints of the selection's linked colours and leaves the rest alone.

{"name":"run_command","arguments":{"command":"paint.setFill","params":{"swatch":"Ink","tint":40}}}
{"name":"run_command","arguments":{"command":"edit.colors.adjustBalance","params":{"mode":"global","tint":-20}}}

Linked gradient stops

Applying a gradient swatch (paint.setFill {swatch}) records it as the gradient's swatch (the Swatches panel highlights it; new stops drop the link). A gradient stop can link to a global or spot swatch like a solid colour: give it swatch (and tint %, default 100 or a tint swatch's own) instead of color in paint.editGradient {stops} or a gradient paint's stops. Editing or deleting the swatch then recolours or unlinks the stop, in art and in gradient swatches. A spot stop separates on its plate; a gradient whose stops are all tints of one spot ink (or paper white, 0 %) exports to PDF as a Separation shading (the writer has no DeviceN, so a gradient mixing a spot ink with other colours is written in process colours, with a warning).

{"name":"run_command","arguments":{"command":"paint.editGradient","params":{"stops":[{"offset":0,"swatch":"Ink"},{"offset":1,"swatch":"Ink","tint":20}]}}}

Transparency flattener presets

flattener.presets.list lists the built-in presets (High, Medium and Low Resolution, builtIn: true) and the saved ones, each with its options. flattener.presets.save {name?, newName?, preset?, …options} creates or changes a saved preset (a change starts from the preset's own options; newName renames it; built-in presets can't change); flattener.presets.delete {name} deletes one. flattener.presets.export {names?, path?} writes them as a .vcflattener JSON file (without path it returns data), and flattener.presets.import {path? | data? | dataBase64?, replace?} adds a file's presets (names in use get a number unless replace). Saved presets live with the preferences and work as preset in object.flattenTransparency (and anywhere else flattener options are taken).

{"name":"run_command","arguments":{"command":"flattener.presets.save","params":{"name":"Press","preset":"high","balance":60}}}
{"name":"run_command","arguments":{"command":"object.flattenTransparency","params":{"preset":"Press"}}}

Flattener Preview

flattener.preview {highlight?, overprints?, preset?, …options, ids?} reports what flattening the document (or ids) would do without changing it: counts of transparent objects, all affected objects, patterns, outlined strokes and type, rasterized complex regions (areas the raster/vector balance rasterizes whole), all rasterized areas and flat-colour regions, plus regions (bounds, and id for objects) for the highlight asked for. overprints: "discard" or "simulate" flatten without preserving overprints.

{"name":"run_command","arguments":{"command":"flattener.preview","params":{"highlight":"allRasterized","preset":"low"}}}

Width profiles

The Profile list holds the built-in variable-width profiles (uniform, lens, taperStart, taperEnd, pinch, teardrop, wave) and profiles saved from strokes; saved ones are kept with the preferences, not in the document (no undo step). stroke.widthProfile.list returns every row (id, label, builtIn, points as [t, left, right]) and current, the selected stroke's profile ("custom" when it isn't listed). stroke.widthProfile.add {name?} saves the selected stroke's variable width (default name "Width Profile N"), stroke.widthProfile.delete {name?} removes a saved one (default: the selected stroke's; built-ins can't be deleted) and stroke.widthProfile.reset removes every saved one. stroke.set {profile} takes a built-in id or a saved name.

{"name":"run_command","arguments":{"command":"stroke.widthProfile.add","params":{"name":"Ribbon"}}}
{"name":"run_command","arguments":{"command":"stroke.set","params":{"ids":[9],"profile":"Ribbon"}}}

Arrowheads

stroke.set {startArrow, endArrow} takes any of 40 generated heads by name (the params doc lists them; null for none): arrows (Arrow, Barbed, Concave, DoubleArrow, HalfArrowLeft/Right, Chevron, Feather, Swallowtail…), filled and open shapes (Triangle, Circle, Oval, Target, Tag, Diamond, Hexagon, Star, the …Open rings…) and marks (Bar, DoubleBar, DotOnBar, Slash, DoubleSlash, Bracket, Fork, Cross, Plus). Each head fits a box about four times its weight (stroke weight × arrowScale) long and wide.

Stroke bounds and clicks

Visual bounds (Fit to Selected Art, Rasterize, exporting all art) and clicks on the canvas take in the whole stroke as drawn: arrowheads, inside/outside alignment (an outside stroke is hit outside the path, an inside one inside it), the width profile's width where you click, projecting caps and miter spikes up to the miter limit. A rectangle's right-angle miters stay within half the weight of its edges.

A symbol instance is measured by its symbol's art, placed by the instance's transform. The bounds in document.inspect and document.node, the Align commands, object.setBounds, the bounding box, clicks, marquees, Fit to Artwork Bounds and exports of the art use that box. So do artboard.rearrange and artboard.move (which artboard the art moves with), artboard.convertToArtboards, select.allOnArtboard, object.transformEach, object.lock.above and object.hide.above. edit.copy and edit.cut record the artboard under the center of that box as the one the copy comes from, and edit.pasteInPlace, edit.pasteInFront, edit.pasteInBack and edit.pasteOnAllArtboards keep the copy's offset to it. edit.paste {center}, library.use {center} and the clipboard.import* commands center that box on the point, and library.add gives a graphic the size of that box. clipboard.exportPng, clipboard.exportPdf and the PNG and PDF that Copy offers other apps are cropped to that box. Visual bounds (Use Preview Bounds, Fit to Selected Art) take in the strokes in the art at their own weight, as the canvas draws them when the instance scales the art. An instance is measured as the 20 × 20 square around the origin, which symbols made in the app keep their art in, when its symbol is missing, when its symbol or a symbol nested in its art holds an instance of itself, and when symbols nest 32 deep in its art. PDF and EPS export and the SVG of an artboard leave out art outside the page or artboard, and Skip Blank Artboards skips an artboard with no art on it. For both, the box of an instance whose art holds brushed strokes grows by 16 times the widest stroke (at least 1 pt) of the objects in its art that have a brushed stroke, plus 8 pt. The SVG of an artboard also keeps an object whose effects paint onto the artboard, and an object that holds an instance when the effects in the symbol's art, or the instance's own, paint onto it.

Expand

object.expand {object?, fill?, stroke?, gradient?, steps?} (Object → Expand…) turns the selection into plain art in one undo step. object (on by default) outlines type, turns live shapes into paths and bakes effects; stroke (on) outlines strokes into filled paths; fill (on) expands gradient fills: with gradient: "objects" (the default) into steps (1–1000, default 255) solid objects (rectangles across a linear gradient, each from its band to the far end so no seams show; concentric ellipses for a radial one, largest first; just the bands when stops are translucent), with gradient: "mesh" into a gradient mesh that paints the gradient (columns or rings where its colour changes, sharp at coincident stops), each inside a clip group shaped like the object. A freeform gradient becomes a mesh shaped like the object either way. Expanded colours keep the colour model their stops share. Expand mixes a freeform gradient's point colors in the model they share to color the mesh points; it mixes C, M, Y and K as inks. When the models differ, the colors mix in display RGB and are stored in the working RGB space, and a CMYK document stores them as CMYK through the color settings. fill also expands pattern fills: each becomes the pattern's tiles that reach the object's box, placed by the pattern's transform, in a clip group shaped like the object that isolates their blending, with the fill's opacity, blend mode and effects; the object's other fills and strokes stay on copies of it in paint order. The tiles' stroke weights and effect distances scale with the pattern, and opacity masks in the tile art, unlinked ones included, move with each tile. When the pattern is rotated, reflected, stretched or sheared, or scaled while its art has geometry effects (Roughen's detail is per inch), the strokes (brushed ones included) and geometry effects of the tiles' paths become filled art and their live shapes become plain paths, in their opacity masks too. A path in the tile art with a Transform effect set to Random, or with a brush that varies at random (a calligraphic or scatter brush with Random variation, or a bristle brush), becomes filled art first, so every tile shows the one result the canvas draws. A pattern fill stays as it is when it lays no tiles there (as when the box would take more than 250,000 tile positions), when a group or type in its art has a Transform effect set to Random, when the pattern is scaled, rotated, reflected, stretched or sheared and its art holds a symbol instance (the canvas scales the instance's strokes with the tile), and when the pattern is rotated, reflected, stretched or sheared, or scaled while its art has geometry effects, and its art has raster or 3D effects, or a geometry effect on a group or type, which the canvas draws turned or scaled with each tile. The geometry effects of a gradient or pattern fill (such as Offset Path) and those of the object shape the outline of the clip group, the gradient fits that outline and the tiles reach its box, as the canvas paints the fill; the expanded art does not keep them. With fill: false gradient and pattern fills stay live. object.expand.info answers which options have something to expand in the selection ({object, fill, stroke, gradient}; gradient says whether Fill has a gradient fill, which Expand Gradient To applies to), and object.mesh.create on a gradient-filled object colours the mesh points as the gradient paints them, with the colors Object › Expand gives them.

{"name":"run_command","arguments":{"command":"object.expand","params":{"stroke":false,"gradient":"mesh"}}}
{"name":"run_command","arguments":{"command":"object.expand","params":{"steps":16}}}

Lab spot colours

Colours can be CIE Lab (D50): give {"l": 55, "a": 60, "b": 40} (L 0–100, a and b about −128–127) wherever a colour is taken; documents store them as {"model": "lab", …}. swatch.new {color, spot: true} or swatch.edit {name, mode: "lab"} defines a spot colour in Lab (Swatch Options' Lab mode). swatch.spotOptions {useLab} sets the document's Spot Colors options: with true (the default) Lab spot colours show and print from their Lab values and PDF export writes their Separation spaces with a Lab alternate; with false art linked to them takes their working-CMYK equivalents and PDF uses a DeviceCMYK alternate. Without useLab it answers the current setting.

{"name":"run_command","arguments":{"command":"swatch.new","params":{"name":"Lab Ink","color":{"l":55,"a":60,"b":40},"spot":true}}}
{"name":"run_command","arguments":{"command":"swatch.spotOptions","params":{"useLab":false}}}

Attributes, URLs and image maps

The Attributes panel (window.panel {panel: "attributes"}, Cmd+F11) reads attributes.info {ids?}: overprintFill, overprintStroke, showCenter, imageMap, url, note, fillRule and reversed (null where the objects differ) and recentUrls. attributes.set {overprintFill?, overprintStroke?, showCenter?, imageMap?, url?, note?, ids?} sets them in one undo step; path.setFillRule {rule: "nonZero"|"evenOdd"} sets the fill rule of the paths and compound paths, and path.reverse {reversed?} makes subpaths run counter-clockwise (true) or clockwise (false). SVG export wraps an object with a URL in <a xlink:href>, and SVG import reads <a href> links back.

{"name":"run_command","arguments":{"command":"attributes.set","params":{"ids":[12],"url":"https://example.com","imageMap":"rectangle"}}}
{"name":"run_command","arguments":{"command":"path.setFillRule","params":{"rule":"evenOdd"}}}

SVG interactivity

The SVG Interactivity panel (window.panel {panel: "svgInteractivity"}) reads svgInteractivity.info {ids?}: events ([{event, script}] in the panel's order, null where the objects differ), files (the linked JavaScript files) and eventNames (the 19 events the panel lists, from onfocusin to onscroll). svgInteractivity.setEvent {event, script, ids?} sets the JavaScript an event runs on the objects, replacing the event's script where they have one (an empty script removes it); svgInteractivity.removeEvent {event, ids?} and svgInteractivity.clearEvents {ids?} remove one event or all of them. Each is one undo step. svgInteractivity.addFile {file} and svgInteractivity.removeFile {file} link and unlink a JavaScript file (a path or a URL, kept as given without surrounding spaces); a document links at most 64 files. ui.javascriptFilesDialog {} opens the JavaScript Files dialog, whose Choose… puts a picked file in as a file:/// URL. SVG export writes each event as an event attribute on the object's element (onclick="…"; with the SVG Tiny 1.2 profile it leaves them out with a warning) and each file as <script xlink:href>. SVG import reads the event attributes the panel lists and the <script> links back as text; it never reads a script file or runs a script, and leaves inline scripts and the links after the first 64 out with a warning. An unnamed <g> with event attributes of its own stays a group, apart from the object in it. The native format keeps the events and the files.

{"name":"run_command","arguments":{"command":"svgInteractivity.setEvent","params":{"ids":[12],"event":"onclick","script":"showMenu()"}}}
{"name":"run_command","arguments":{"command":"svgInteractivity.addFile","params":{"file":"menu.js"}}}

Removing anchor points

path.removeAnchors {} (Object › Path › Remove Anchor Points) removes the direct-selected anchors (select.anchors {id, anchors: [[subpath, anchor]…], mode?}) without opening their paths, in one undo step, and answers {removedObjects} (paths left with no segment go). Each removed point's neighbours keep their handle directions and their facing handles are refitted so one cubic follows the two old segments; two straight sides become one. The anchor after each removed one is then selected (the one before at an open path's end), so running path.removeAnchors again goes on along the path. path.removeAnchor {id, subpath?, anchor} (the Delete Anchor Point tool) removes one anchor the same way, and path.deleteAnchors {} (the Delete key) deletes the selected anchors with their segments, opening closed paths there.

{"name":"run_command","arguments":{"command":"select.anchors","params":{"id":12,"anchors":[[0,1],[0,3]]}}}
{"name":"run_command","arguments":{"command":"path.removeAnchors","params":{}}}

Converting and cutting anchor points

The Control bar's and the Properties panel's anchor buttons, one undo step each, act on the direct-selected anchors (all anchors of a path selected as a whole): path.convertAnchors {to: "corner"|"smooth"} retracts both handles, or pulls handles out in line with the neighbouring anchors (a smooth anchor keeps its own); path.cutAtAnchors {} (Cut Path at Selected Anchor Points) cuts there and answers {ids}: a closed path opens at the cut, its two ends on top of each other, and an open path becomes one path per piece. Each cut leaves one of its two anchors selected, so path.moveAnchors {dx, dy} (or a Direct Selection drag) pulls the path apart there; path.join {} (Connect Selected End Points) joins the ends again; path.join {ids: [a, b], ends: ["last", "first"]} joins the ends asked for (the Pen's join, #776: drawing on from one open path, a click on another's end makes them one path). path.convertAnchor {id, subpath?, anchor, to, x?, y?} and path.split {id, subpath?, anchor} do the same to one anchor (the Anchor Point and Scissors tools); the Pen with Alt held over a selected path's handle, anchor or segment works as the Anchor Point tool. path.reshapeSegment {id, subpath?, segment, t, dx, dy, perpendicular?} (a segment dragged with Direct Selection or the Anchor Point tool) moves the segment's point at t by dx, dy: a smooth anchor at either end stays smooth, its other handle turning with the moved one. With perpendicular: true (Shift held during the drag) both of the segment's handles stand at right angles to the line between its anchors, on the same side and at the same length, and the point at t lies as far from that line as the point moved by dx, dy does; a segment whose anchors coincide bends as it does without perpendicular. path.reshapeSegment refuses a delta of 1e9 pt or longer ("invalid delta"), as object.blend.spine.reshapeSegment does. path.convertAnchor (its x, y) and path.freehand (each of its points) refuse a point whose x or y is 1e9 pt or more in size ("x and y must be finite"). The canvas check after an edit measures a segment of a selected path by its control points when one of them has an x or y that is infinite or larger in size than f64::MAX / 16 (about 1.1e307), so an edit that leaves such a path selected, path.create included, fails with "result would exceed the canvas (coordinates out of range)". path.curvatureEdit, path.removeAnchor, path.convertAnchor, path.reshapeSegment, path.split, path.setHandle, path.moveAnchor and path.insertAnchor edit subpath 0 when subpath is absent or null, and refuse a subpath that is not a whole number of 0 or more, such as 1.5, -1 or "1" ("missing integer subpath"); path.split refuses such an anchor or segment in the same way. The document does not change and no undo step is recorded.

A pointer_gesture with mods: {cmd: true} held at the press, with any tool but the selection tools (Command on macOS, Ctrl elsewhere), drags with the selection tool chosen last (select_tool), or Direct Selection with the Pen, Curvature and anchor tools until one is chosen; the release gives the tool back as it was, so the Pen goes on drawing the path it had begun while that stays selected. Its reply's tool is the tool given back.

{"name":"pointer_gesture","arguments":{"tool":"pen","events":[{"kind":"down","x":100,"y":300},{"kind":"up","x":100,"y":300},{"kind":"down","x":200,"y":300},{"kind":"drag","x":250,"y":300},{"kind":"up","x":250,"y":300}]}}
{"name":"pointer_gesture","arguments":{"mods":{"cmd":true},"events":[{"kind":"down","x":250,"y":300},{"kind":"drag","x":250,"y":260},{"kind":"up","x":250,"y":260}]}}
{"name":"pointer_gesture","arguments":{"events":[{"kind":"down","x":300,"y":350},{"kind":"up","x":300,"y":350}]}}

{"name":"run_command","arguments":{"command":"select.anchors","params":{"id":12,"anchors":[[0,2]]}}}
{"name":"run_command","arguments":{"command":"path.cutAtAnchors","params":{}}}
{"name":"run_command","arguments":{"command":"path.moveAnchors","params":{"dx":0,"dy":40}}}

Simplify

object.path.simplify takes the amount as tolerance (points, 1 by default) or as the Simplify slider's level, a whole number from 0 (the fewest anchor points) to 100 (the closest fit). Level 50 is a tolerance of 1 pt, and every 10 levels halve or double it, so level 0 is 32 pt and level 100 is 1/32 pt. auto: true (Auto-Simplify) uses the lowest level whose tolerance is at most 1/400 of the selection's longer side, or level 100 when that side is under 12.5 pt. With level or auto, the reply also has the level used:

{"name":"run_command","arguments":{"command":"object.path.simplify","params":{"auto":true}}}

In the desktop app, Object › Path › Simplify opens the slider (ui.menuDialog {command: "object.path.simplify"}, dialog simplifySlider with field level) and previews the result until ui.dialog.confirm. ui.simplifyAuto moves the open slider to the Auto-Simplify level, and ui.simplifyMoreOptions opens the Simplify dialog at the slider's tolerance. When OK confirms the dialog with retain ticked, the menu item opens the dialog on those settings next time.

Editing a path with the Curvature tool

The Curvature tool edits any selected path, whatever drew it, keeping its shape except where edited (#798). Each edit is one undo step and leaves the anchor it edited direct-selected (the tool's current point): path.curvatureEdit {id, subpath?, op, …} answers {anchor}. op: "move" with anchor, x, y moves a point and re-curves only its two segments: a smooth point takes the tool's handles (along the circle through it and its neighbours; the one handle into the curve at an end), and the neighbours' handles keep their directions (their lengths scale with the chord), so the segments beyond don't change. op: "insert" with segment, t adds a smooth point there without changing the shape (with x, y it is then moved there: the tool's click-and-drag on a segment). op: "extend" with x, y goes on from the end (default) or the start of an open subpath (the new end curves into the path unless corner is true), and op: "close" closes it from that end; from says how the old end bends into the new segment: "keep" (default) keeps its curve, "smooth" curves it through its neighbours (a point the tool placed), "corner" leaves it without handles. Alt-click or double-click toggles a point with path.convertAnchor, and Backspace/Delete removes the current point with path.removeAnchor.

{"name":"run_command","arguments":{"command":"path.curvatureEdit","params":{"id":12,"op":"move","anchor":2,"x":260,"y":140}}}
{"name":"run_command","arguments":{"command":"path.curvatureEdit","params":{"id":12,"op":"extend","end":"end","x":550,"y":120}}}
{"name":"run_command","arguments":{"command":"path.curvatureEdit","params":{"id":12,"op":"extend","x":650,"y":60,"from":"smooth"}}}

Registration and trim marks

Every document has the built-in [Registration] swatch (listed after None by swatch.list): a colour that prints on every plate, process and spot. paint.setStroke {swatch: "[Registration]"} applies it; it can't be edited, moved, duplicated, merged or deleted. Separations Preview shows it on each plate and PDF export writes it as /Separation /All. object.createTrimMarks {style?, allArtboards?} draws trim marks in Registration around the selection, or around every artboard when nothing is selected; effect.apply {effect: "cropMarks"} adds live crop marks that follow the object. Both use Japanese marks (double lines at the trim and bleed edges, centre marks) when the preference japaneseCropMarks is on or style: "japanese" is given.

{"name":"run_command","arguments":{"command":"prefs.set","params":{"key":"japaneseCropMarks","value":true}}}
{"name":"run_command","arguments":{"command":"object.createTrimMarks","params":{}}}

PDF export draws printer's marks and bleed itself (the Save PDF dialog's Marks and Bleeds section): document.exportPdf {marks: {trim, registration, colorBars, pageInfo, kind, weight, offset}, bleed: {useDocument, top, bottom, left, right}}. Each page's TrimBox is its artboard, its BleedBox the artboard grown by the bleed (the document's with useDocument, set by document.setup {bleed}; the art there is kept) and its MediaBox the bleed box grown to hold the marks (the art is clipped to the bleed box). Trim marks, registration targets and the page information (title, artboard, export date in UTC) are drawn in Registration (/Separation /All) outside the bleed; the colour bars are process, spot and black-tint patches. Layers whose Print option is off (layer.setProps {printable: false}) are left out unless includeNonPrinting (or createLayers) is on.

{"name":"run_command","arguments":{"command":"document.exportPdf","params":{"path":"/tmp/press.pdf","preset":"Press Quality","marks":{"trim":true,"registration":true,"colorBars":true,"pageInfo":true}}}}

The Output and Advanced sections convert colours and write real text. output: {conversion, destination} converts every colour into the destination profile's model (destination; CMYK of another profile too) or only the colours of the other model (preserveNumbers: colours already in it keep their numbers); grey stays grey, images are converted too and printer's marks are not. The destination is any profile edit.colorSettings lists (default: the document's for its colour mode). profiles (all, destination, or taggedSource for a document assigned profiles with edit.assignProfile) writes colours in ICC-based spaces with their profiles embedded: CMYK with the destination's or the document's CMYK profile, RGB as sRGB, grey with the sRGB tone curve (PDF/A, PDF/X-3 and PDF/X-4 files are always tagged, whatever profiles says). outputIntent embeds a profile as the file's /GTS_PDFX output intent with outputCondition, outputConditionId and registry, and trapped sets /Trapped (PDF/A files keep their own output intent). Type is written as selectable, searchable text in embedded subset fonts with a ToUnicode map (fonts whose licence forbids embedding stay outlines, with a warning); advanced: {outlineText: true} writes every glyph as an outline instead. Type that Transform effects only move and copy (no scale, rotation, reflection or Random) is written as text once in each place when its characters are in solid colors without underline or strikethrough, the type object has no fills or strokes of its own, no inline graphics and no perspective, its transform does not scale or shear stroked characters, and the copies do not overlap. When the effect is on the type itself, the type can also have transparency, an opacity mask and raster effects such as a drop shadow. When the effect is on the group or layer holding the type, the type has no transparency or effects of its own and is not in a clipping path, the group or layer has no fill, stroke or Pathfinder effect of its own, and no group between them has a fill, stroke, Pathfinder or geometry effect. Type is written as outlines when the effect stops its copies early, or when the copies would number more than 10,000 or take more than 20,000,000 bytes of document data. Other type under a geometry effect is written as outlines.

{"name":"run_command","arguments":{"command":"document.exportPdf","params":{"path":"/tmp/press.pdf","output":{"conversion":"preserveNumbers","destination":"VectorCraft Generic CMYK (SWOP-like)","profiles":"destination","outputIntent":"VectorCraft Generic CMYK (SWOP-like)","trapped":true},"advanced":{"outlineText":false}}}}

Scale Strokes & Effects

object.scale, object.transform and object.transformEach take strokes?: bool (Scale Strokes & Effects) and corners?: bool (Scale Corners); without them the preferences scaleStrokes and scaleCorners (both off by default) apply (prefs.set {key, value}). With strokes on, a scale by k (the square root of the transform's determinant) multiplies stroke weights, dash lengths and dash offsets, and every distance parameter of the object's, its fills' and its strokes' effects (drop shadow offsets and blur, offsets, radii…; effect.list gives each effect's lengths; relative Roughen, Tweak and Zig Zag sizes stay percentages), in groups and layers too. With it off, nothing painted changes size, and type keeps its character strokes' weight. With Scale Corners off, live corner radii keep their size. The journal entry of a scaling command records the strokes and corners it used, so actions replay alike. object.transformAgain applies the last transform's matrix with the copy, strokes, corners, patterns and typeAreas values that transform used, and a true or false value given in its own params takes precedence. After a drag of art without a pattern, patterns follows the preference transformPatternTiles. Its journal entry records the strokes, corners and patterns it used, as a transform's does.

{"name":"run_command","arguments":{"command":"object.scale","params":{"sx":200,"strokes":true}}}
{"name":"run_command","arguments":{"command":"object.transformEach","params":{"scaleH":50,"scaleV":50,"strokes":false}}}

Live Corners

object.setLiveShape {id?, ids?, radius?, kind?, corners?, items?} sets the corners of any path (one undo step): a live rectangle's or polygon's, a star's, a pen path's. A corner is an anchor where two sides, straight or curved, meet at an angle (not an open path's ends, not a smooth anchor, not one the sides run straight on through); a curved side is trimmed to what the cut leaves of it. radius (pt) and kind (round, invertedRound or chamfer) go to the corners given, else to the corners holding a Direct-Selected anchor (select.anchors), else to every corner. corners are anchor indices of the path with its corners uncut, counting every subpath's anchors in order: a rectangle's 0 top-left, 1 top-right, 2 bottom-right, 3 bottom-left; a polygon's from its first vertex clockwise; a star's from its first tip; an index past the last anchor is an error. Each corner keeps its own radius and kind (the shape's live in queries has radii and, when a corner isn't round, kinds); a corner with no radius is one anchor, a cut one two, and Direct-Selected corners stay selected as that changes. A path that isn't a live shape keeps its uncut outline (live {"shape": "path", "base", "radii"}) so its corners stay editable, and is a plain path again once no corner is cut; a polygon stays a live polygon, its corners keeping the radius they shared when sides changes. Every corner is a circular arc tangent to both sides: a radius draws no larger than takes the cut halfway along the corner's shorter side (half the shorter side of a rectangle, the same limit for its four corners), the corners stay circular through uneven scales (the radius scales by the mean scale, or keeps its size with Scale Corners off), and a rectangle from a file that kept an uneven scale in its transform (elliptical corners) gets circular ones in document units when its corners are next set. Dragging a corner widget rounds the corners whose widgets show (every corner, or the Direct-Selected ones) on every selected path at once, outlining in red those that reach their limit; the Selection tool shows the widgets of live rectangles and polygons, the Direct Selection tool those of any path. Several objects each with their own corners are one call: items: [{id, corners?}] in place of ids and corners. Alt-clicking one cycles their kind and double-clicking one opens Corners (ui.corners {id?, corners?}, dialog corners: kind, radius; OK runs object.setLiveShape).

{"name":"run_command","arguments":{"command":"object.setLiveShape","params":{"id":12,"corners":[1],"radius":16}}}
{"name":"run_command","arguments":{"command":"object.setLiveShape","params":{"id":12,"corners":[0,3],"radius":8,"kind":"chamfer"}}}
{"name":"run_command","arguments":{"command":"shape.star","params":{"cx":200,"cy":200,"radius1":60,"radius2":30}}}
{"name":"run_command","arguments":{"command":"object.setLiveShape","params":{"radius":5}}}

The same command sets the rest of a live shape's properties, as the Properties and Transform panels show them. A polygon (Polygon Properties): sides (3–1000), polygonAngle (degrees counterclockwise, its first vertex straight up at 0), polygonRadius (pt, centre to vertex), sideLength (pt, which sets the radius) and makeSidesEqual: true, which drops an uneven scale or shear (its sides then differ) and keeps its centre, angle and mean radius; a radius or length that isn't a number above 0 is an error. An ellipse (Ellipse Properties): pieStart and pieEnd (degrees counterclockwise from 3 o'clock; 0 to 360 is the whole ellipse) and invertPie: true, which swaps them. Each call is one undo step.

{"name":"run_command","arguments":{"command":"object.setLiveShape","params":{"sides":8,"polygonAngle":22.5,"sideLength":40}}}
{"name":"run_command","arguments":{"command":"object.setLiveShape","params":{"makeSidesEqual":true}}}

Align to Artboard

object.align {to: "artboard", artboard?} aligns the selected objects to artboard artboard (a 0-based index). Without artboard, or with an index the document doesn't have, it aligns them to the artboard under the center of the selection, or to the first artboard when no artboard is under the center. With to: "selection" or "key", object.align does not use artboard. select.allOnArtboard {artboard?} (Select › All on Active Artboard) selects the objects on artboard artboard, or on the first artboard when the call gives none; an index the document doesn't have is an error.

object.distribute {to: "artboard", artboard?} and object.distributeSpacing {to: "artboard", artboard?} (the Distribute Objects and Distribute Spacing buttons with Align to Artboard) use the same artboard as object.align. object.distribute moves the first object, in the order of the line it spaces (left edges, centers or right edges; top edges, centers or bottom edges), until its left or top edge is on the artboard's left or top edge, and it moves the last object until its right or bottom edge is on the artboard's right or bottom edge. It spaces the other objects evenly by that line between them, and two objects are enough. object.distributeSpacing without spacing puts the left or top edge of the first object, in the order of those edges, on the artboard's left or top edge, and it puts the right or bottom edge of the last object on the artboard's right or bottom edge, with equal gaps between the objects. With to: "selection" or "key", or without to, object.distribute and object.distributeSpacing without spacing keep the first and last objects where they are, and object.distribute without spacing moves nothing with fewer than three objects. With spacing and a key object, object.distribute and object.distributeSpacing do not use to; with a key guide, to decides whether the guide is the key. With anchor points selected (Direct Selection), both commands space the points among themselves and do not use to or artboard.

object.distribute {spacing} (the Distribute Objects buttons while the selection has a key object, select.key) keeps the key object where it is, with the compound path or selected group that contains it. It places the objects' lines (left edges, centers or right edges; top edges, centers or bottom edges) spacing points apart, outward from the key object's line, in their order on that line. With spacing: 0, every object ends on the key object's line. Two objects are enough. Without a key object or key guide, or with anchor points selected, object.distribute does not use spacing. A spacing that is not a number is an error.

object.distribute and object.distributeSpacing space the selected ruler guides across the axis (vertical guides for the horizontal options, horizontal guides for the vertical ones) with the selected objects, as items with no size (#1597). A guide's left edge, center and right edge are all its line, so the three options of an axis give the same result for it. Guides along the axis stay where they are. Two or more guides selected without objects are enough, and the Align panel's buttons are enabled for them; with to: "artboard", the outermost items go to the artboard's edges.

With a key guide (select.key {guide}, #2143), object.distribute {spacing} and object.distributeSpacing {spacing} (the Align panel's Distribute buttons send the Distribute Spacing value then too) keep the guide where it is and space the other items from it, as from a key object with no size: object.distribute places their lines spacing points apart, outward from the guide's line, and object.distributeSpacing leaves gaps of spacing points, the guide included. With spacing: 0, object.distribute puts every line on the guide. A key guide along the axis leaves everything where it is, as object.align does. to: "selection" or "artboard" distributes without the key guide as the key, as with object.align, and without spacing a key guide is spaced as any selected guide.

The desktop app adds its active artboard, the one the Artboards panel highlights, to the object.align, object.distribute, object.distributeSpacing and select.allOnArtboard calls it runs without artboard, as it does for edit.pasteInPlace, edit.pasteInFront and edit.pasteInBack, and to each command.batch. A batch passes its artboard on to each of these steps that names none, when the step runs. When the document a step runs in has no artboard at that index (for example, after an earlier step deleted the active artboard when it was the last one), the step gets that document's last artboard. In remote mode, MCP's run_command calls and each step of the command_batch tool get the active artboard too. The Actions panel plays an action as a command.batch. The Actions panel menu's Batch… runs action.batch: the same steps on every readable file in a folder (or every data set of the open document), with destination none | saveAndClose | folder and Override Open / Save / Export. Over data sets, the folder destination and the overrides name each file after the document and the data set (<document>_<data set>.<ext>). Batch writes /, \, :, <, >, ", |, ?, * and control characters in the data set's name as -, and adds -2, -3… to a name an earlier data set used (compared without case). Caps: 500 files or data sets, 8 folder levels, 256 steps. An action recorded in the Actions panel leaves the artboard out, and its steps of these commands get the artboard that is active when the action plays. A press with the Selection, Direct Selection or Group Selection tool inside an artboard makes that artboard the active one; with the Selection tool, a Shift-click on an empty spot inside an artboard makes it active and keeps the selection. Choosing the Artboard tool (tool.select {tool: "artboard"}) starts it on the active artboard: its active option is that artboard's index.

artboard.fitToArt {index? | indices?} (Object › Artboards › Fit to Artwork Bounds) fits each artboard it names to the art on that artboard, in one undo step. The art is what an export draws, measured with its strokes and with the effects on paths, compound paths and symbol instances. An object counts for every artboard its bounds overlap, also when it reaches past the artboard's edge, and a group or a clipping layer counts whole. An artboard with no art on it keeps its size and place, and the call fails with no artwork when none of its artboards has art. artboard.fitToSelection {index?} (Object › Artboards › Fit to Selected Art) fits artboard index to the selected art. Both commands take 0-based indexes and resize the first artboard when the call names none. The desktop app adds the selected artboards (those the Artboard tool has selected, else the active one) as indices to the artboard.fitToArt calls it runs without index or indices. It adds its active artboard as index to the artboard.fitToSelection calls it runs without index. Both include MCP's run_command calls in remote mode. The Artboards panel menu's Fit to Artwork Bounds also passes the selected artboards. A batch passes its artboard on to each of their steps that names no index, as the step's index. When the document a step runs in has no artboard at that index, the step gets that document's last artboard. An action recorded in the Actions panel leaves index and indices out of these steps, and they resize the artboard that is active when the action plays.

{"name":"run_command","arguments":{"command":"object.align","params":{"horizontal":"left","to":"artboard","artboard":1}}}
{"name":"run_command","arguments":{"command":"object.distribute","params":{"horizontal":"left","to":"artboard","artboard":1}}}
{"name":"run_command","arguments":{"command":"select.allOnArtboard","params":{"artboard":1}}}

Use Preview Bounds

With the preference usePreviewBounds on (prefs.set {key: "usePreviewBounds", value: true}; also the Align panel flyout's Use Preview Bounds), the Transform panel's and Control bar's X, Y, W and H, the bounding box, and object.align, object.distribute and object.distributeSpacing measure visual bounds, which take in the whole stroke (see Stroke bounds and clicks); off, they measure the paths. The three Align commands also take bounds: "preview"|"geometric" for one call. object.setBounds then sets the visual box: with Scale Strokes & Effects off a 100 pt wide rectangle with a 10 pt stroke set to width: 220 gets a 210 pt path.

{"name":"run_command","arguments":{"command":"object.align","params":{"horizontal":"left","bounds":"preview"}}}

object.align needs horizontal or vertical. The three Align commands, object.reflect and object.shear return an error that lists the choices of a horizontal, vertical, to, axis or bounds param given another value (object.reflect also takes an angle and names the value given), and one for a spacing that is not a number; nothing moves.

Selection preferences

The Selection & Anchor Display and General preferences apply to pointer_gesture as they do to the mouse:

• selectionTolerance (1–8 px, 3 by default): how near a click must be to a path to pick it, and to an anchor, handle or segment for Direct Selection, the Anchor Point, Add and Delete Anchor Point and Scissors tools, and the Pen with Alt. A handle shorter than that leaves its anchor to be picked when pressed nearer the anchor.
• objectSelectionByPathOnly: a click inside a filled path or compound path doesn't select it, one on its path does.
• ctrlClickSelectsBehind (on by default): a Selection tool click with mods: {cmd: true} (Command on macOS, Ctrl elsewhere) selects the object under the selected one there, the next such click the one under that, then the topmost again. Cmd held to borrow the selection tool from another tool clicks as usual.
• doubleClickToIsolate (on by default): a doubleclick with the Selection tool isolates a group, a compound path or a single path, image, mesh or placed file (not type, blends, envelopes, repeats or symbol instances); off, it no longer does.
• usePreciseCursors: the Pen, Eyedropper, Slice and Blend tools' pointers are a crosshair.
• snapToPointTolerance (1–8 px, 2 by default): with View › Snap to Point on and Smart Guides off (desktop app), the point a selection is dragged by, a drawn point and a transform tool's reference point land on an anchor or a ruler guide that near.
• moveLockedWithArtboard: artboard.move {moveArt: true}, the Artboard tool and artboard.rearrange move locked and hidden art with the artboard too, and artboard.setProps {scaleArt: true} scales it too; off (the default) it stays where it is.
• penRubberBand, curvatureRubberBand (on by default): off, the Pen and Curvature tools draw no preview segment to the pointer.
• showHandlesMultipleAnchors (on by default): off, Direct Selection (and the Anchor Point tool, and the Pen with Alt) shows and drags direction handles only while a single anchor is selected. On, the handles of the direct-selected anchors, or of every anchor of a path selected as a whole, show and drag. handleStyle (solid, hollow, large) draws their ends (desktop app).
• hideCornerWidgetAbove (177° by default): corners wider than this show no Live Corners widget (a rectangle's right angles hide below 90°).
• transformPatternTiles (off by default): the default of the transforms' patterns param (object.transform, object.move, object.rotate, object.scale, object.reflect, object.shear, object.transformEach, object.align, object.distribute, object.distributeSpacing, the Selection and transform tools, the dialogs' Transform Patterns) and of the same param on the commands that move or duplicate art with an artboard (artboard.move and the Artboard tool, artboard.duplicate, artboard.rearrange, artboard.setProps): pattern fills and strokes move with the art. object.nudge (the arrow keys) and path.moveAnchors (a Direct Selection drag of selected anchors) take the same param, with the same default, for the objects they move whole.
• selectSameTintPercent (off by default): select.same.fillColor, strokeColor and fillAndStroke take every tint of a global or spot swatch; on, only the same tint.

{"name":"run_command","arguments":{"command":"prefs.set","params":{"key":"objectSelectionByPathOnly","value":true}}}
{"name":"pointer_gesture","arguments":{"tool":"selection","mods":{"cmd":true},"events":[{"kind":"down","x":175,"y":125},{"kind":"up","x":175,"y":125}]}}

The preferences for the view need the desktop app (vectorcraft-cli mcp --connect): zoomWithMouseWheel (the wheel zooms about the pointer, Shift-wheel scrolls up and down, Cmd/Ctrl-wheel sideways; the control channel's ui.wheel turns the wheel), trackpadRotateView (on by default: a two-finger rotation on a Mac trackpad, in the macOS app or in Safari, turns the view about the pointer once it reaches 5°; with it off, the gesture does not turn the view), zoomToSelection (on: Zoom In and Zoom Out centre the selection), showToolTips, anchorSize (1–7), gridColor, gridStyle, gridsInBack, guideColor, guideStyle, showPixelGrid (on by default: in View › Pixel Preview at 600% zoom and above, a line at every document pixel over the art), recentFontsCount, antiAliasedArtwork (on by default; off, the canvas draws the art with hard edges — a pixel is painted when the art covers at least half of it — while raster effects and pattern tiles stay smooth; exports keep their own Anti-aliasing option), scrubNumericFields (on: a horizontal drag on a numeric field's label steps the field, one undo step per drag; the control channel's ui.drag scrubs), showHomeScreen (on by default: the Home screen while no document is open; off, an empty window, and the Home button or app.home still shows the screen), autoCollapseIconPanels (off by default: on, a click away from a panel popped out of the icon column puts it away), displayPrintSize (off by default: View › Actual Size / view.actualSize is one document point per screen point; on, one document inch fills 96 screen points, an inch at the system's reference density whatever the display scaling). highlightAlternateGlyphs (off by default: on, the canvas paints a pink cell behind glyphs OpenType features substituted — ligatures, contextual forms — on screen; Document Setup's Highlight Substituted Fonts / Glyphs stay separate). blackOnScreen (Appearance of Black › On Screen, accurate by default) sets how the canvas shows pure CMYK black, black ink alone at 100%: accurate shows it through the CMYK profile, a dark grey, and rich shows it as RGB black. Rich blacks, black tints and images show the same either way, and exports are unchanged. blackOutput (Appearance of Black › Printing / Exporting, accurate by default) does the same for RGB raster exports (document.export, document.serialize, document.exportSelection, document.exportForScreens, document.exportForOffice) and for Print as Bitmap: rich draws pure CMYK black as RGB black in the pixels; CMYK ink exports and the document's colours stay as they are. antiAliasedBitmaps (off by default; desktop app): in View › Pixel Preview, on softens placed and embedded bitmaps; off, their pixels stay hard-edged. disableAutoAddDelete (off by default): on, the Pen no longer adds or deletes an anchor when a click lands on a segment or an anchor of a selected path.

gridlineEvery (72 pt by default) and gridSubdivisions (8 by default) set the document grid that View › Show Grid draws, and View › Snap to Grid snaps to its subdivisions. Changing either sets the grid of every open document. file.new gives a new document that grid, and document.open, file.newFromTemplate and file.revert give it to every document they read except a native file and the native document that an SVG, PDF, .ai or EPS file carries (restored: true), which keep the grid they were saved with.

view.rotate {angle} (desktop app) turns the view to angle degrees about the center of the canvas; a positive angle turns the art clockwise on screen. View › Rotate View and the status bar's angle menu run it. It returns {angle}, the angle it set, normalized to (-180, 180], and view.rotateReset turns the view back to 0. view.rotateToSelection (desktop app, View › Rotate View to Selection) turns the view about the center of the canvas to the angle the selected objects share (selectionRotation in document.inspect, 0 when their angles differ), so their bounding box is square to the window. It returns {angle}, and with nothing selected it returns "nothing selected".

The Smart Guides preferences (Preferences › Smart Guides) apply to pointer_gesture with Smart Guides on (the default view) and to the mouse; they change what the tools show and how far a target pulls, and only Construction Guides change where a snapped point lands:

• smartGuideColor (#ff3dfc by default): the colour of the smart guides' lines and labels.
• alignmentGuides (on by default): off, no line is drawn along the edge or centre the art lines up with; the art still snaps into line.
• anchorPathLabels (on by default): off, no "anchor", "center", "path" or "align" label.
• measurementLabels (on by default): off, no size or offset readout while drawing, moving (Free Transform's move too), dragging a ruler guide, drawing and resizing an artboard, or moving and resizing a slice.
• transformToolsGuides (on by default): off, no readout while scaling or rotating with the Selection tool, the Free Transform tool or the Rotate, Scale, Shear and Reflect tools.
• objectHighlighting (on by default; desktop app): off, the Selection tools no longer outline the object under the pointer. The outline also needs View › Smart Guides on.
• snappingTolerance (0–40 px, 4 by default): how near an anchor, edge, centre or artboard edge pulls a drawn point, a dragged selection, a bounding-box handle, a ruler guide or an artboard. With Smart Guides off, Snap to Point uses snapToPointTolerance instead.
• constructionGuides (on by default) and constructionAngles (90° & 45° Angles by default; also 90° Angles, 45° Angles, 60° Angles, 30° Angles, 90° & 45° & 30° Angles): a point drawn near a line at one of those angles through the point it leaves from (the Pen's or Curvature tool's last anchor, a line's start) lands on that line, where a target's alignment line crosses it if one does nearby. "90°" gives the horizontal and the vertical, "45°" the diagonals, "60°" and "30°" the multiples of those angles in between. Off, the point only lines up.
• spacingGuides (on by default): a moved selection (Selection tool / pointer_gesture with selection) snaps so the gap to a neighbour matches the equal spacing among other objects nearby; off, only alignment and construction guides pull.

With Smart Guides on, a bounding-box handle dragged with the Selection or Free Transform tool (mods.shift proportional, mods.alt from the centre) lands on another object's anchor or centre (labelled "anchor" or "center"), else its corners and side line up with the edges and centres of the other objects and the artboards (a line and "align"). The Rotate, Scale, Shear and Reflect tools start a drag from the anchor or centre they grab and snap the pointer as drawing tools do, so dragging a corner onto another object's anchor scales exactly onto it:

{"name":"pointer_gesture","arguments":{"tool":"scale","events":[{"kind":"down","x":100,"y":100},{"kind":"up","x":100,"y":100},{"kind":"down","x":199,"y":198},{"kind":"drag","x":302,"y":262},{"kind":"up","x":302,"y":262}]}}

Every point a drawing tool places snaps the same way, through one snapper (#506): the Pen's and Curvature tool's anchors, the Pencil's first and last point, the shape tools' start and dragged corner or radius (rectangle, roundedRectangle, ellipse, polygon, star, lineSegment, arc, spiral, rectangularGrid, polarGrid), the Type tool's click and frame, and a new artboard's corners. The point lands on another object's anchor, centre or path ("anchor", "center", "path"), else lines up with the anchors, edges and centres of the art and the artboards (a line from the target and "align"); the Pen's lands on the path being drawn too, so a click on its first anchor closes it exactly. With mods.shift the point keeps to its 45° step from the anchor before it (from constrainAngle; a box dragged keeps its corner on a diagonal) and slides along that way into line with the nearest target. A move event before the press shows where it would go, as hovering does with the mouse. The targets are the art in the window (all of it headless), gathered once per document state; the shape being drawn is never one of them:

{"name":"pointer_gesture","arguments":{"tool":"rectangle","events":[{"kind":"down","x":202,"y":203},{"kind":"drag","x":330,"y":341},{"kind":"up","x":330,"y":341}]}}
{"name":"pointer_gesture","arguments":{"tool":"pen","events":[{"kind":"down","x":300,"y":420},{"kind":"up","x":300,"y":420},{"kind":"down","x":151.5,"y":431},{"kind":"up","x":151.5,"y":431}]}}

With a 100 × 100 square at (100, 100), the first starts the rectangle on its corner (200, 200); the second puts the Pen's second anchor at (150, 431), in line with the square's centre.

Type preferences

• placeholderText (on by default): type the Type tools place (pointer_gesture with type, areaType…) starts with placeholder text, selected, so type_text replaces it; text.create and text.createInPath take placeholder: true.
• typeSizeIncrement, trackingIncrement, baselineShiftIncrement: what type.step {attribute: "size" | "leading" | "tracking" | "kerning" | "baselineShift", by?} steps by, on the text range given, the type objects in ids, the Type tool's selected text or the selected type objects. type.size.increase / type.size.decrease (Cmd+Shift+. and Cmd+Shift+,) step the size. type.size.increaseByFive (Cmd+Alt+Shift+., press_key {key: ".", mods: {cmd: true, alt: true, shift: true}}) and type.size.decreaseByFive (Cmd+Alt+Shift+,) step it by five times typeSizeIncrement. The Type tool's Alt+arrows (press_key {key: "Right", mods: {alt: true}}) step kerning at a caret or tracking of the selection (←/→), leading (↑/↓) and baseline shift (Shift+↑/↓), five steps with Cmd/Ctrl too. type.resetTracking (Cmd+Alt+Q, press_key {key: "Q", mods: {cmd: true, alt: true}}) sets tracking to 0 and kerning to Auto in the type objects in ids, the Type tool's selected text or the selected type objects; at a caret it resets only the kerning of the character before it. type.resetHorizontalScale (Cmd+Shift+X, press_key {key: "X", mods: {cmd: true, shift: true}}) sets the horizontal scale to 100%, as the Character panel shows it, in the same text; at a caret it leaves the text as it is.
• missingGlyphProtection (on by default): text.setStyle and text.setRangeStyle changing the font leave the characters the new font has no glyph for in the font that had one.
• typeSelectionByPathOnly (off by default): on, the Selection tools (pointer_gesture with selection, and the mouse) pick type on its type path only — the baseline of each of its lines (a vertical column's centre line), area type's frame, type on a path's path — within selectionTolerance; a click among the glyphs, between the lines or inside the frame selects nothing. Off, anywhere in the type's bounds selects it. Marquee selection is unchanged, and the Type tools (a click among the characters edits them) and the Eyedropper (eyedropper) are not affected.
• fontNamesInEnglish (on by default): Preferences › Type › Show Font Names in English. Off, the Character panel and Type → Font menus label a family by its native-language name when the font has one. text.fontList and text.setStyle {font} always use the English (canonical) family name; a document may still name a font by any of its aliases (ヒラギノ角ゴシック resolves to Hiragino Sans).
• hyphenationExceptions (empty by default): Preferences › Hyphenation › Exceptions. A comma-separated list of words, matched without regard to case: a word with no hyphens is never hyphenated (typography), and a word with hyphens breaks only where they are (hap-pen). prefs.set pushes the list into the hyphenator and reflows open documents.
• highlightAlternateGlyphs (off by default): Preferences › Type › Highlight Alternate Glyphs. On, the canvas paints pink behind OpenType substitutions (ligatures, contextual forms). Document Setup's Highlight Substituted Fonts / Glyphs stay separate. Agents set it with prefs.set; the desktop canvas reads it on the next paint.

{"name":"run_command","arguments":{"command":"prefs.set","params":{"values":{"typeSizeIncrement":4,"trackingIncrement":50}}}}
{"name":"run_command","arguments":{"command":"prefs.set","params":{"key":"fontNamesInEnglish","value":false}}}
{"name":"run_command","arguments":{"command":"prefs.set","params":{"key":"hyphenationExceptions","value":"typography, hap-pen"}}}
{"name":"run_command","arguments":{"command":"prefs.set","params":{"key":"highlightAlternateGlyphs","value":true}}}
{"name":"run_command","arguments":{"command":"type.step","params":{"attribute":"tracking","by":-2}}}

File Handling › appendConverted (on by default): an older native file (the former .drawcraft extension or an older format version) opens titled <name> [Converted], and Save doesn't write over it: document.save without a path answers with a suggested .vectorcraft name (the app asks for one). Off, it opens under its own name and Save writes over it.

Width points

stroke.widthPoint.set {id, t, left, right, index?, adjustAdjoining?} adds a width point (side widths in points) or, with index, edits or moves one; adjustAdjoining changes the nearest points either side in proportion. Moving a point onto another one's t (or stroke.widthPoint.copy {id, index, t} landing there) makes a discontinuous point: two points at the same t (the width before it, then after it), and the stroke's width steps there. document.inspect reports a path's strokeOptions.widthPoints ([t, left, right], factors of half the weight; null for a uniform stroke). stroke.widthPoint.remove {id, index | indices} deletes points; stroke.widthProfile.set {ids, points} replaces them all, as dragging several Shift-selected points with the Width tool does in one undo step. With the Width tool, press_key Delete or Backspace removes the selected width points (handledBy: "tool"); with none selected the key clears the selected objects as usual (edit.clear). The Puppet Warp tool's selected pins go the same way. A compound path's stroke is the compound's: width point commands given one of its members edit the compound's profile, which runs along every subpath (the Width tool shows each subpath's points).

object.puppetWarp {ids?, pins, moved, angles?, expand?} warps art as rigidly as possible so that each pin (a point on the art) lands on its moved point. angles (degrees, or null for a free pin; as long as pins) holds the turn the art takes around a pin, as Alt-dragging near a selected pin with the Puppet Warp tool does: one pin with angles: [90] turns the art a quarter turn around it.

The Puppet Warp tool's pins live in the document while the same art stays selected (never saved; Undo and Redo take them back with the art). object.puppetWarp.pins {ids?} reads them: pins (where each sits on the shape the pins started from), moved (where it is now), angles, and auto: true while none was placed (the tool's automatic pins: the centre and the end of each limb). Edit moved, angles or the lists and pass the result back to object.puppetWarp: with rest: true the warp always starts from that rest shape, so warps don't stack and putting a pin back restores the original; every pin must be on the art's mesh, and empty lists remove the pins.

{"name":"run_command","arguments":{"command":"object.puppetWarp.pins","params":{}}}
{"name":"run_command","arguments":{"command":"object.puppetWarp","params":{"rest":true,"pins":[[250,150],[110,150],[390,150]],"moved":[[250,150],[110,150],[390,90]],"angles":[null,null,null]}}}

{"name":"run_command","arguments":{"command":"stroke.widthPoint.set","params":{"id":9,"t":0.5,"left":4,"right":4}}}
{"name":"run_command","arguments":{"command":"stroke.widthPoint.set","params":{"id":9,"t":0.8,"left":12,"right":12}}}
{"name":"run_command","arguments":{"command":"stroke.widthPoint.set","params":{"id":9,"index":1,"t":0.8,"left":4,"right":4}}}

New art

Initially (and after paint.default) new art has a white fill and a 1 pt black stroke. New objects take the fill and stroke proxies (paint.setFill / paint.setStroke and the weight, whatever is selected) on top of a template. With nothing selected, stroke.set / stroke.setAdvanced (cap, join, dashes, arrowheads, profile…) and graphicStyle.apply (every fill, stroke and effect, opacity, blend mode and the link to the style; add: true stacks it) set that template instead of editing art. With the Appearance panel's New Art Has Basic Appearance off (appearance.setNewArtBasic {on: false}, the preference newArtBasic) new art takes the whole appearance of the last selected object. appearance.newArt reports what the next object gets; paint.default resets it.

{"name":"run_command","arguments":{"command":"select.none","params":{}}}
{"name":"run_command","arguments":{"command":"stroke.set","params":{"cap":"round","dash":[6,3]}}}
{"name":"run_command","arguments":{"command":"appearance.newArt","params":{}}}

Gradients on strokes

A linear or radial gradient on a stroke lies within it (placed on the page like a fill's gradient: the default), along it (from the start of each subpath to its end) or across it (from the stroke's left edge to its right edge, left of the path's direction; an inside or outside stroke spans the side it shows on). paint.editGradient {stroke: true, strokeMode: "within"|"along"|"across"} sets it (the Gradient panel's Stroke buttons), and document.inspect reports it as strokeOptions.gradientMode. Width profiles, dashes and arrowheads keep working: the gradient runs on through gaps and into the heads. Outline Stroke and Expand turn such a stroke into gradient meshes clipped to its outline; SVG and PDF write it as slices of linear gradients clipped to the outline (vectorcraft_svg::export_with_report and the PDF report warn about it).

{"name":"run_command","arguments":{"command":"paint.editGradient","params":{"stroke":true,"strokeMode":"along"}}}

Units

Commands take and return lengths in points (the distances of object.path.offsetPath, object.path.simplify, object.path.splitIntoGrid and text.setStyle size/leading also take a string with a unit, "5 mm"). A length string takes the same arithmetic as the desktop app's fields, where % is a share of the value before it: "200mm*50%" is 100 mm and "50pt+25%" is 62.5 pt. The units only change what the UI shows and how it reads typed numbers. Preferences ▸ Units has three: General (rulers, positions and sizes, the Info panel, dialog distances, canvas measurement labels), Stroke (weights, dashes) and Type (font size, leading, baseline shift). General is the active document's units (document.inspect → units): document.setUnits {units} (Document Setup) sets it for that document, and prefs.set {key: "unitsGeneral", value} sets it for the active document too (one undo step) and is the units file.new starts in when it gets no units. Stroke and Type are the preferences unitsStroke and unitsType. numbersWithoutUnitsArePoints (on by default) makes the desktop app's length fields in picas read a number typed with no unit in points (12 is 12 pt, 2p6 and 1p stay picas, 12 mm millimetres); off, it is in picas. Fields in other units read it in their unit either way, and the Preferences dialog dims the option unless a unit is Picas. Command params are in points either way. identifyObjectsBy (objectName by default, or xmlId) sets how Layers, Variables and Layer Options show and edit object names: Object Name keeps free text; XML ID shows AI-style XML ids (My Rect #2 as My_Rect__x23_2) and layer.setProps / rename require a valid XML ID (letter, underscore or colon first; no spaces), stored back as the object name. prefs.list marks the length preferences (keyboardIncrement, cornerRadius, pasteOffset, gridlineEvery, typeSizeIncrement, baselineShiftIncrement) with measure: "general"|"type"; they are kept in points and also take a string with a unit. Unit names: Points, Picas, Inches, Millimeters, Centimeters, Pixels, Feet & Inches, Meters, Yards, Feet (the preferences use points, picas, … feetInches, meters, yards, feet).

{"name":"run_command","arguments":{"command":"prefs.set","params":{"key":"unitsGeneral","value":"millimeters"}}}

Document Setup

document.setup with no params reports the document setup: units, the bleed ([top, bottom, left, right] in pt, drawn as a red outline around each artboard), the transparency grid (size and two colours; the first is also the simulated paper colour), the flattener preset and Discard White Overprint (Overprint Preview keeps white overprints visible while it is on), the type options (language and its quotes, Use Typographer's Quotes for typed quotes, superscript, subscript and small caps proportions, SVG text export) and the background contents (white: raster exports are white behind the art). Pass any of these keys to change them in one undo step; document.setUnits stays as an alias for units. A file whose bleed (0 to 72 pt a side), superscript and subscript sizes (1 to 500%) and positions (-500 to 500%) or small caps size (1 to 200%) lies outside those ranges opens with the nearer end of each range.

{"name":"run_command","arguments":{"command":"document.setup","params":{"gridColors":"Blue","gridSize":"large","language":"German","bleed":9}}}

New Document

file.new takes everything New Document sets: preset (a name from file.newPresets), name, width/height (points, or lengths such as "210 mm"), units, orientation, artboards with artboardLayout {layout, columns, spacing, rightToLeft}, bleed, backgroundContents, colorMode, rasterEffectsPpi (1 to 2,400; a saved preset's value outside that range is read as the nearer end) and previewMode. file.newPresets {category?} lists the categories and presets (Recent: the last sizes used; Saved: the user's presets, kept in the preferences by file.newPresets.save and removed by file.newPresets.delete). A listed preset can be passed straight back to file.new. Print presets and sizes without units start in unitsGeneral; screen presets (mobile, web, video, social) in Pixels. A saved or Recent preset whose size (up to 163,830 pt), artboards (1 to 1,000), spacing (0 to 100,000 pt) or bleed (0 to 72 pt a side) lies outside those ranges is read with the nearer end of each range, and a size of 0 or less is read as 1 pt. The desktop app's file.newDialog (Cmd/Ctrl+N) opens the New Document dialog on the first Recent entry, or on the Print category's Letter while Recent is empty. file.newWithoutDialog (Cmd/Ctrl+Alt+N) runs file.new with those settings and the next untitled name, without opening the dialog, and returns what file.new returns.

{"name":"run_command","arguments":{"command":"file.new","params":{"preset":"A4","orientation":"landscape","artboards":4,"artboardLayout":{"columns":2},"bleed":9}}}

Place

file.place puts another file's art into the active document as one undo step without touching the clipboard: a raster image at 100% of its physical size (the resolution its file declares, else 72 ppi; linked to its path unless link: false; a Photoshop document as its merged image), an SVG as one group (fitted by its art's bounds; with crop: "crop" given, framed and clipped by its viewBox, #788), a PDF/.ai page as one clipped group, with the images, symbols, patterns and swatches it uses. A VectorCraft document read from path is a placed document: one locked object showing its artboard page (or, with crop: "bounding", its art's bounds), linked to the file and read again when it changes (see Linked images), and vectors in every output; with link: false (or from dataBase64) it is an editable copy of its art, as one clipped group. at centres it, rect fits it, replace swaps the selected object (keeping its place and transform), template puts it on a new template layer. The art's bounds are the bounds of the art that shows, with its strokes and effects, as an export measures them: hidden objects, guides and template layers do not count, and symbol instances are measured by their art. A page or artboard holds the objects whose shown art lies on it; a hidden object on it is still placed and stays hidden, as do the hidden members of the groups placed. replace and the width and height that file.place answers also measure symbol instances by their art. file.place.info describes a file without placing it and image.info reports a placed image's link, colour mode and effective ppi. file.place.queue loads the place cursor (the place tool) with several files: headless, drive it with pointer_gesture (a click places at 100% with the top-left corner there, a drag at the dragged size) and press_key (Left/Right/Up/Down cycle, Escape discards the current file).

{"name":"run_command","arguments":{"command":"file.place","params":{"path":"/tmp/photo.jpg","at":[300,200]}}}
{"name":"run_command","arguments":{"command":"file.place","params":{"name":"logo.svg","dataBase64":"PHN2Zy…","rect":[0,0,100,100]}}}
{"name":"run_command","arguments":{"command":"file.place.queue","params":{"paths":["/tmp/a.png","/tmp/b.pdf"]}}}
{"name":"pointer_gesture","arguments":{"events":[{"kind":"down","x":40,"y":40},{"kind":"up","x":40,"y":40}]}}

Copy and paste between documents

edit.copy keeps the copied objects with the document resources they use: image blobs, symbols, patterns, global and spot swatches (with the tint swatches of the tints used; never the built-in [Registration] swatch), the gradient swatches their gradients came from, graphic styles the objects are linked to, character and paragraph styles and brushes. Width profiles live in the preferences, so they need no copying. Every edit.paste* command (also edit.pasteWithoutFormatting, which leaves text styles behind) brings them into the active document in the same undo step: an identical resource is reused, a missing one added, and one of the same name that differs is added renamed ("Mark 2"; a copy an earlier paste added is reused), which the pasted objects follow. Linked colours show the document's swatch at their tint (a Lab spot colour as its Spot Colors option says). The result reports {ids, added, merged, renamed: [{kind, from, to}]}.

A global or spot swatch whose name the document gives another colour is a conflict: clipboard.conflicts lists them, and swatchConflict answers: "merge" (the default: the objects take the document's swatch), "add" (the pasted swatch comes in renamed) or one answer per name. Pasting back into the document the objects came from uses its resources as they are now and raises no conflict.

Placement: edit.paste {center} centres the objects on a point (the app passes the view centre), else offsets them by dx/dy (the Paste Offset preference). With nothing selected, edit.pasteInFront / edit.pasteInBack put them on top / at the bottom of the current layer. edit.pasteOnAllArtboards keeps their offset to the artboard they were copied from. layer.pasteRemembersLayers {on?} (a document option, document.inspect → pasteRemembersLayers) pastes objects back into the layers they came from, by name, making missing ones.

Artboards copy with their art. artboard.copy {index? | indices?, art?} puts artboards on the clipboard together with the art fully inside each (index and indices default to the artboards the Artboard tool has selected; art defaults to the tool's moveArt option, Move/Copy Artwork with Artboard; locked and hidden art only with prefs moveLockedWithArtboard); with the Artboard tool chosen, edit.copy does this for the tool's selected artboards and edit.cut runs artboard.cut, which also deletes them (one artboard is always left: a cut of every artboard fails). Then edit.paste adds a copy of each artboard, together right of the last one and placed as they lay, with their art in the layers it came from, in one undo step (edit.pasteInPlace, edit.pasteInFront and edit.pasteInBack put them where they were; edit.pasteOnAllArtboards pastes only the art), in this document or another; the result's artboards lists the new artboards' indices and artboard is the first of them, and the Artboard tool, when chosen, selects them. artboard.duplicate {index? | indices?, art?} does the same in one step → {index, indices, ids}. Window › Artboards › Duplicate Artboards passes the selected artboards (those the Artboard tool has selected, else the active one) as indices, and a row dragged onto New Artboard passes its artboard as index. artboard.move {copy: true, moveArt: true} is the Artboard tool's Alt-drag. artboard.duplicate and artboard.move also move the pattern tiles of the copies when patterns is true (default: prefs transformPatternTiles); edit.paste leaves the tiles of a pasted artboard's art where they were. Every artboard command that takes an index fails with no such artboard when the index is not the 0-based index of an artboard (past the last one, negative, fractional, a string or null), and the document and the clipboard stay as they were.

Scale Artwork with Artboard (#602): artboard.setProps {index, x?, y?, width?, height?, scaleArt: true} resizes an artboard and scales the art fully inside it, and its artboard guides, from the old rectangle onto the new one, each side by its own ratio (keep the proportions for an even scale), in one undo step → {scaled: [id…]}. Locked and hidden art scales only with lockedAndHidden: true (default: prefs moveLockedWithArtboard); strokes, effects and corners scale as with object.scale (strokes?, corners?, default: Scale Strokes & Effects and Scale Corners), pattern tiles with patterns? (default: Transform Pattern Tiles). A move that keeps the size scales nothing (art moves with artboard.move {moveArt}). The Artboard tool's scaleArt option (tool.setOption {tool: "artboard", key: "scaleArt", value: true}; the Control bar's and Properties' Scale Artwork with Artboard check box, beside Move Artwork with Artboard) makes its handle drags proportional and passes scaleArt, as do Properties' and the Transform panel's artboard size fields while the tool is in use.

artboard.setProps {artboards: [{index, name?, x?, y?, width?, height?}…], scaleArt?, moveArt?, …} sets several artboards in one undo step, each with its own fields; the other params apply to all of them. The art that moves or scales with each artboard is the art inside it before any of them changes, and an object inside several of them goes with the first one listed that moves or scales its art. An empty list and a list that names an artboard twice are refused.

With several artboards selected with the Artboard tool, the desktop app gives a new size to each of them in one such call. Properties' W and H and the Transform panel's W and H do this, and so do Properties' Preset and orientation buttons, the Control bar's preset, orientation, W and H, and a new Width or Height in Artboard Options. The W and H of the Transform panel and the Control bar resize each artboard about its own reference point, and with the proportions linked they keep each artboard's own proportions. Artboard Options resizes each artboard about the reference point its form used for the new Width or Height. Properties' W and H and the preset and orientation buttons keep each artboard's top-left corner. A preset sizes each artboard in its own orientation, and an orientation swaps the width and height of each artboard that is not square and not already in it. A new X or Y moves the active artboard only, and Artboard Options' Name renames only the artboard it shows.

{"name":"run_command","arguments":{"command":"artboard.setProps","params":{"artboards":[{"index":0,"width":300},{"index":2,"width":300}],"scaleArt":true}}}

{"name":"run_command","arguments":{"command":"artboard.copy","params":{"index":0}}}
{"name":"run_command","arguments":{"command":"edit.paste","params":{}}}
{"name":"run_command","arguments":{"command":"artboard.duplicate","params":{"index":0,"art":true}}}
{"name":"run_command","arguments":{"command":"clipboard.conflicts","params":{}}}
{"name":"run_command","arguments":{"command":"edit.paste","params":{"center":[300,200],"swatchConflict":{"Brand":"add"}}}}
{"name":"run_command","arguments":{"command":"layer.pasteRemembersLayers","params":{"on":true}}}

File Info

file.info with no params reports the document's File Info: title, author, authorTitle, description, keywords, rating (0–5), copyrightStatus (unknown, copyrighted, publicDomain), copyrightNotice, copyrightUrl, and the read-only created (set by file.new) and modified (set by every save to a file) dates as ISO 8601 UTC. Pass any of the editable keys to change them in one undo step (null clears one; keywords takes a list or a comma-separated string and keeps each word once); bad values are refused. PDF export writes the title, author, description and keywords to the document info and XMP, PNG export writes them (with the copyright and creation date) as text chunks, and SVG export writes the description as <desc> and, with the metadata option, all of it as Dublin Core. PDF export and PDF-compatible .ai and .ait saves write the created date (the time of writing when the document has none) as the file's creation date and the time of writing as its modification date. A PDF, .ai or .ait opened from its pages or its editing copy (document.open, file.newFromTemplate, file.revert; not one restored from the document it carries) opens with the author, description, keywords, copyrightStatus, copyrightNotice, copyrightUrl, created and modified that its XMP packet (dc:creator, dc:description, dc:subject or pdf:Keywords, xmpRights:Marked, dc:rights, xmpRights:WebStatement, xmp:CreateDate, xmp:ModifyDate) and its document information dictionary (Author, Subject, Keywords, CreationDate, ModDate) give, the XMP packet's value where both give one. A value that file.info would refuse is left out, and the title is set as for any other open (the file name, or a new untitled name for a template). file.new {created} and a save to a file (document.save or file.saveAs with {path, modified}) use the given date (Unix seconds, or null for none) instead of now. Their journal entries always record the date they used, so replaying the journal makes the same document whenever it runs (an action recorded in the Actions panel leaves the date out: playing it later dates the document then). As a step of command.batch, a command records the date it used (and any value it takes from the preferences) in its step of the batch's journal entry. A batch may open, switch, close or revert documents: each document it edits gets one undo step, the batch is journaled whichever document it ends in, and an error rolls every document back (the ones it opened close, the ones it closed or reverted come back). The rollback puts the documents back, not the disk: files its steps wrote (a save, a Save As, an export) stay written, new files included, as Undo leaves a saved file on disk. A caller that wants nothing written unless the whole batch succeeds puts the saves and exports last, or runs them after the batch.

{"name":"run_command","arguments":{"command":"file.info","params":{"author":"Ada","keywords":"poster, fair","rating":4,"copyrightStatus":"copyrighted","copyrightNotice":"© 2026 Ada"}}}

Document Raster Effects Settings

document.rasterEffectsSettings holds how raster effects (shadows, glows, blurs, feathers) become images when PDF export or Expand Appearance renders them: resolution (1 to 2,400 ppi, also New Document's rasterEffectsPpi; a file's value outside that range opens as the nearer end), colorModel (the document's rgb/cmyk, grayscale or bitmap), background (white makes the images opaque), antiAlias (off: hard edges), clippingMask (the white stays under the art only), addAround (0 to 1,000 points of room around the art; a file's value outside that range opens as the nearer end) and preserveSpotColors (stored). They are also the defaults of object.rasterize, whose params override them (its clippingMask puts the image in a clip group with the art's outline). No params reports them; any of them changes them in one undo step. PDF export renders the raster effects of every kind of object this way (paths, groups, layers, type, images, symbol instances, live objects, and those in opacity masks, symbols and pattern tiles): shadows and outer glows go in an image under the untouched vector art, other effects replace the object with its image, and a path whose fills or strokes carry effects is written as one piece per fill and stroke, so only the affected ones become images. When an image in the art can't be decoded, object.rasterize and object.cropImage (without trim) make their image without it, and their results carry the warning that Flatten Transparency gives.

{"name":"run_command","arguments":{"command":"document.rasterEffectsSettings","params":{"resolution":"high","background":"white","addAround":36}}}

effect.apply {effect: "rasterize"} (Effect › Rasterize…) draws an object as an image of square pixels while its paths, type and other art stay editable. It takes the options of Object › Rasterize under these names: colorModel (document, grayscale or bitmap), useDocumentResolution (true by default: the resolution above) or resolution (1 to 2400 ppi), background (transparent by default, or white), antiAlias (art, type or none), clippingMask and addAround (points). The image covers the object's bounds and the room around them, its first pixel at their top-left corner. Both object.rasterize and the effect measure the art as drawn: with its geometry effects, its brush art and its raster effects. The effects above Rasterize in the stack, shadows included, are part of the image, and the shadows and glows below it are cast by the image. The canvas draws an image of up to 8.4 million pixels, at most 32,768 a side, on its own grid and then at the view's zoom: an image pixel wider than a screen pixel shows as one flat square, and smaller ones are averaged. A larger image is made from the art drawn at the canvas's resolution (lower for very large art), and its pixels no wider than a pixel of that drawing show the art as drawn. When Rasterize is an object's only raster effect, Expand Appearance and PDF and SVG export replace the object with the image object.rasterize makes with the same options.

{"name":"run_command","arguments":{"command":"effect.apply","params":{"effect":"rasterize","params":{"useDocumentResolution":false,"resolution":150,"background":"white"}}}}

Glass textures

Glass (distort.glass) takes an image file as its surface in params.textureImage, as {path} or {name?, dataBase64}: a PNG, JPEG, GIF, WebP, TIFF, BMP, Targa, PSD or PSB file, decoded within the pixel allowance the renderer draws images in, counted as RGBA. effect.apply and effect.setParams store the image's gray levels in the effect as a grayscale PNG in base64 and set texture: "loaded" unless the params give a texture. effect.list reports the stored string, and a command given that string stores it unchanged. Any other string, a file that is not an image, or an image past the allowance is an error. A save keeps the texture with the effect. Glass lays the texture one image pixel per point at Scaling 100%, centered on the object's center and repeated in both directions. Each pixel's luma is the surface's height, from 0 for black to 1 pt for white. A transparent pixel counts as middle gray, and a partly transparent pixel is mixed with middle gray by its alpha.

{"name":"run_command","arguments":{"command":"effect.apply","params":{"effect":"distort.glass","params":{"distortion":10,"textureImage":{"path":"/tmp/texture.png"}}}}}

Linked images

A placed image linked to its file (file.place {path}, Link on) records the file's absolute path, its size, modification time and hash, and keeps a low-resolution preview (at most 256 px a side): a .vectorcraft save writes the preview instead of the pixels, plus the path relative to the saved file. document.open reads the linked files again, looking for each at its path, then at its relative path and by name in the document's folder (so a folder moved with its links still opens). Only files on this computer, or on the network share the document is on, are looked at: a path to another computer, a device or a URL counts as missing. On Windows, a document opened from a mapped drive (Z:\…) is on the share the drive is mapped to, and links on that share (\\server\share\…) are looked at. A link on a drive mapped to another share (Y:\…) counts as missing, from a document on this computer too. The result lists missingLinks (their images show the preview), modifiedLinks (left as they were; read again only with the preference updateLinks: "automatically", then they are in updatedLinks), each as {name, path, ids}, with refused (why the path isn't followed) when it isn't. The desktop app also checks the active document's links when its window comes back to the front, following the same preference. links.check reports every link's status (ok, modified, missing), links.update {ids?} reads modified files again and links.relink {ids?, path | folder} points images at another file (or each at the file of its name in a folder); images keep their bounds, one undo step each. Without a file system (the web), linked images show their previews.

links.findFiles {folder} looks in an absolute folder and its subfolders, on separate threads, for the files of the active document's missing links, or of the images and placed documents ids whatever their status, and returns at once with an id and the search's state (searching while it runs). It finds the files with each linked file's name, in any letter case, and lists them best first as {path, size, match, caseDiffers?}: same when the file has the size and hash saved with the link, size when it has the saved size and the link saved no hash (or the file is larger than 1 GB, which a search doesn't hash), else name. Within each kind, a file whose name has the link's letter case comes first. Then each file's folder names are compared with the saved path's, from the innermost outward up to the first one that differs, and a file with more matching names comes first; files that still tie are listed by path. links.findFiles {} returns the last search's state, and {stop: true} stops it. A search looks for at most 1,000 linked files, keeps the first 10 files of each one's name and up to 10 that are the same file, and ends once every linked file has its same file; it searches only the folders text.findFontFiles may search, with the same maxSeconds, depth and count limits. links.relink {ids, path, copy: true} copies the file into the saved document's folder and links images to the copy, never replacing a file there: the names Name.png, Name 2.png up to Name 99.png are tried in turn until one is free, which gets the copy (copied), or is a file with the same contents, which is used (kept). With copy, a call that includes a placed document fails and relinks nothing; placed documents are relinked without copy, where their files are. Find in Folder… in the Linked File Not Found dialog and in the Links panel menu (ui.findLinksInFolder) runs both. The web build has no folder search.

Placed documents (a .vectorcraft file placed linked) work the same way: the document keeps the file's bytes and a preview (JPEG or PNG, 1 px per point, at most 1024 px a side), and a save writes the preview (Include Linked Files: the file); output reads the file again, or uses the preview with a warning when it is gone or changed. links.list and links.info give them document: true and the artboard's pageWidth and pageHeight (pt). Saving the file in the app updates the open documents that place it. links.embed (the Links panel's Break Link) turns one into the editable copy placing without link gives, its symbols, patterns, swatches and images joining the document; so does object.expand.

{"name":"run_command","arguments":{"command":"links.check","params":{}}}
{"name":"run_command","arguments":{"command":"links.relink","params":{"ids":[12],"path":"/new/photo.png"}}}
{"name":"run_command","arguments":{"command":"links.update","params":{}}}
{"name":"run_command","arguments":{"command":"links.findFiles","params":{"folder":"/Users/me/Archive"}}}
{"name":"run_command","arguments":{"command":"links.findFiles","params":{}}}
{"name":"run_command","arguments":{"command":"links.relink","params":{"ids":[12],"path":"/Users/me/Archive/2024/photo.png","copy":true}}}

Placing a text file (.txt) sets it as area type: file.place {path | name+dataBase64, text?: {characterSet?: "unicode" | "ansi", platform?: "windows" | "mac", removeLineReturns?, removeParagraphReturns?, replaceSpaces?: n}} (Text Import Options). Unicode reads UTF-8, or UTF-16 with a byte-order mark; ANSI reads Windows-1252 (Mac Roman on mac). removeLineReturns joins the lines of each block into one paragraph, removeParagraphReturns drops blank lines, replaceSpaces: 3 turns runs of 3 or more spaces into tabs. The frame fills rect, the replaced object's bounds, or the artboard less a 36 pt margin.

{"name":"run_command","arguments":{"command":"file.place","params":{"path":"/tmp/notes.txt","text":{"removeLineReturns":true,"removeParagraphReturns":true}}}}

An RTF file (any file but a .txt that starts with {\rtf) is placed the same way, without text options, and keeps its formatting: each run's font, size, bold and italic, underline, strikethrough, all caps, small caps, superscript or subscript, baseline shift and color, and each paragraph's alignment, indents, space before and after and direction. Table cells are separated by tabs and each row is a paragraph; hidden text, pictures, headers, footers and notes are left out. file.place answers format: "rtf".

A Word document (a ZIP package named .docx, or one that holds word/document.xml, except a .txt file) is placed the same way and keeps the same formatting, taken from the document defaults, the paragraph style and the character style (and the styles each is based on) and direct formatting; theme fonts are read from the theme part. Table cells are separated by tabs and each row is a paragraph. Hidden and deleted text, field instructions, pictures, text boxes, headers, footers, notes and list numbers are left out. A package with more than 10,000 files, a part that unpacks to more than 16 MiB or nests elements more than 256 deep, and a document with more than 4 MiB of text or 100,000 paragraphs or runs are refused. file.place answers format: "docx".

PDF presets

pdf.preset.list lists the built-in presets (VectorCraft Default, which preserves editing, then High Quality Print, Press Quality, Smallest File Size and the PDF/X presets, builtIn: true; supported says the writer produces the preset's standard) and the saved ones, each with its description and settings. pdf.preset.save {name?, newName?, description?, preset?, …document.exportPdf options} creates or changes a saved preset (a change starts from the preset's own settings; newName renames it; built-in presets are read-only; passwords are never stored); pdf.preset.delete {name} deletes one. pdf.preset.export {names?, path?} writes them as a .vcpdfpresets JSON file (without path it returns data), and pdf.preset.import {path? | data? | dataBase64?, replace?} adds a file's presets (names in use get a number unless replace). Saved presets work as preset in document.exportPdf, document.pdfSettings, document.export/serialize with format pdf and .ai saves. They are kept across restarts, in the desktop app and in vectorcraft-cli alike (see What a headless session remembers).

{"name":"run_command","arguments":{"command":"pdf.preset.save","params":{"name":"Web","preset":"Smallest File Size","compatibility":"1.5"}}}
{"name":"run_command","arguments":{"command":"document.exportPdf","params":{"preset":"Web","path":"/tmp/web.pdf"}}}

Clipboard formats

Besides SVG, Copy offers other apps PNG and PDF of the copied objects, and their text when they are all type. clipboard.flavours lists what Copy offers for the current clipboard, best first, by the Clipboard Handling preferences: text/plain (the text of a type-only copy, else the SVG markup with copyAsSvg when the images it embeds total at most 1 MiB, so a paste never reads a photo back as text, #2213), image/svg+xml (copyAsSvg), application/pdf (copyAsPdf) and image/png (always; PDF and PNG need objects with an area). Each one has a query command: clipboard.exportSvg, clipboard.exportText (null unless every copied object is type), clipboard.exportPdf (a one-page PDF, the page the objects' visual bounds) and clipboard.exportPng {scale?} (cropped to the objects, transparent around them). As in document.export, the PDF's and the PNG's replies list their notes in warnings (type written in the fallback font…; the PNG's also say when an image couldn't be decoded). After Copy or Cut the app shows the first note of the PDF and PNG it offers in the status bar.

What other apps copied goes into the clipboard with clipboard.importSvg, clipboard.importPdf {dataBase64, page?, password?, editingData?} (the page's objects; with editingData: true the data is a .ai or .ait file, and every object of its editing data comes in as document.open reads it, each layer's objects in stacking order and a hidden layer's objects hidden, or the page's objects when document.open would not use that copy), clipboard.importEps {dataBase64} (an EPS or other PostScript picture's objects, as document.open reads one, a hidden layer's objects hidden), clipboard.importImage {dataBase64, mime?} (an embedded image at 100% of its physical size) or clipboard.importText {text} (point text in the default type style), each centred on center (default: the first artboard); then any edit.paste* command pastes it. The desktop app does this itself: Copy and Cut publish the formats to the system clipboard, and a Paste outside text editing first reads what another app copied (SVG, then PDF, then text, then a bitmap; SVG markup in text counts as SVG). A background thread makes the formats and they are published when they are ready. While that thread has a save, an export, a print or a recovery copy to finish, Copy and Cut make and publish them at once. When another Copy or Cut, or a copy of text in VectorCraft, comes first, they are dropped. Pasting what VectorCraft copied keeps the lossless internal clipboard, also before its formats are published. While a text field or the Type tool has the keyboard, the app waits for them. Windows carries every format both ways and pastes a bitmap from PNG, else from the device-independent bitmap screenshots copy (CF_DIBV5, then CF_DIB; alpha kept, opaque when it is zero throughout, at most 32768 px a side); macOS and Linux carry one format (the text, else the PNG) and paste text and bitmaps (macOS the pasteboard's PNG, else TIFF, as screenshots copy them). On the web Copy publishes SVG text, and Paste takes SVG text and the pictures and files a paste carries (a screenshot pastes as an embedded image). A .ai or .ait file that holds a PDF, copied in a file manager or carried by a paste, is loaded with clipboard.importPdf {editingData: true}, and its objects go into the active layer.

The pasteTextFormatting preference (keep or plain): with plain, text the Type tool copied pastes into type without its formatting, taking the style at the caret.

{"name":"run_command","arguments":{"command":"clipboard.flavours","params":{}}}
{"name":"run_command","arguments":{"command":"clipboard.exportPng","params":{"scale":2}}}
{"name":"run_command","arguments":{"command":"clipboard.importImage","params":{"dataBase64":"iVBORw0KGgo…","center":[300,200]}}}
{"name":"run_command","arguments":{"command":"edit.paste","params":{"center":[300,200]}}}

DXF export

document.exportDxf (also document.export / export with format dxf) writes a CAD drawing, ASCII DXF R12 to 2018 (version: R12, R13, R14, 2000, 2004, 2007, 2010, 2013, 2018; default 2018). Each layer becomes a DXF layer (hidden layers switched off, locked ones locked, non-printing ones not plotted; template layers left out), straight paths become polylines, curved ones cubic splines through every anchor, fills their outlines followed by solid hatches (laser and cutter software reads the outlines and skips hatches; R12 has no hatches: the outlines alone), strokes lines with their lineweight and a linetype per dash pattern, and placed images image entities linked to PNG or JPEG files (rasterFormat) written next to the drawing (linked in the result). Coordinates are y up from the bottom-left corner of the first artboard (or artboard), in drawing units: scale units per unit (default 1 mm = 1 unit, which sets $INSUNITS); scaleLineweights scales the lineweights with them. colors is 8, 16 or 256 indexed colours, or true (default; true colour with the nearest index, DXF 2004 and later). preserve: "appearance" (default) writes type as glyph outlines and the strokes a CAD line can't draw (inside or outside, width profiles, arrowheads) as filled outlines; "editability" keeps type as text and every stroke a line. alterPaths writes every stroke as its filled outline, outlineText outlines type, selectedOnly writes only the selected objects in their layers (every export takes it), and useArtboards: true writes one drawing per chosen artboard holding the art over it. What DXF can't hold (gradients and patterns as one colour, blending, opacity masks, raster effects, clipping) comes back in warnings.

DWG can't be written (no openly licensed writer exists): document.formats lists it under unsupported with that hint, and exporting to it answers with the hint to export DXF. PICT files are not supported either and say so when opened.

{"name":"run_command","arguments":{"command":"document.exportDxf","params":{"path":"/tmp/plan.dxf","version":"2013","unit":"mm","scale":10,"preserve":"editability"}}}
{"name":"export","arguments":{"path":"/tmp/plan.dxf","options":{"useArtboards":true,"colors":256}}}

DXF import

document.open reads ASCII DXF drawings of any version (by content or the .dxf extension), and file.place places one as a group. Both take dxf: {layout, unit, scale, fit, fitTo, scaleLineweights, center, mergeLayers} (all optional). document.dxfInfo reads a drawing without opening it: its version, drawing units, layouts (Model first, then the paper layouts), layers, and the default ratio (unit, scale: the drawing at 1:1 in its own unit; unitless metric drawings read millimetres, imperial ones inches). A ratio is 1 unit of the art = scale drawing units; fit: true instead scales the art to fit fitTo (a letter page when opening; when placing, the artboard under at). center: false puts the drawing's origin (fitted art: its bottom-left corner) on the artboard's bottom-left corner; mergeLayers puts all art on one layer. Each DXF layer holding art becomes a layer (off and frozen layers hidden, locked ones locked, non-plotting ones non-printing). Lines, polylines (arc segments from bulges), circles, arcs, ellipses, splines (NURBS, within a millionth of their size), solids and hatches become paths (solid hatches filled, pattern hatches as their boundaries), text and multiline text point type, named blocks symbols (one per look when their art takes colour "by block"), anonymous blocks such as dimensions groups. Indexed colours (7 is black on paper), true colour, lineweights (scaleLineweights scales them with the art), linetypes and transparency resolve through layers and blocks. What is left out (viewports, images, meshes, unknown entities) is listed in warnings. Binary DXF and DWG files answer an error saying to save them as ASCII DXF.

{"name":"run_command","arguments":{"command":"document.dxfInfo","params":{"path":"/tmp/plan.dxf"}}}
{"name":"run_command","arguments":{"command":"document.open","params":{"path":"/tmp/plan.dxf","dxf":{"layout":"Layout1","fit":true,"mergeLayers":true}}}}
{"name":"run_command","arguments":{"command":"file.place","params":{"path":"/tmp/detail.dxf","dxf":{"unit":"mm","scale":10,"center":false}}}}

Links panel

links.list {show?: all|missing|modified|embedded, sort?: name|kind|status} lists every image in the layers, top first, as the Links panel does: {id, name, linked, status: ok|modified|missing|embedded, format, pixelWidth, pixelHeight, path?, found?, page?, preview?, note?} (note: the image is too large to decode and isn't drawn; document.open and file.place say the same in their warnings). links.info {id?} is the Link Info: image.info's fields plus status, format, the file's ppi and the effectivePpi, scale (% of 100%), rotation (degrees, counter-clockwise), placement and, for a linked file, fileName, location, fileSize, created, modified. links.goTo {id} selects an image; links.embed {ids?} keeps linked files' pixels in the document (an image whose file is missing stays linked: relink it first); links.unembed {id, path} writes an embedded image to a file (another extension converts it) and links to it, and without path returns {name, dataBase64}. links.placementOptions {ids?, preserve?: transforms|bounds|fileDimensions|fit|fill, align?: topLeft…bottomRight, clip?} decides how a relinked or updated file takes an image's place (default bounds: stretched into the old bounds; clip puts it in a clip group of the old bounds when it is larger). The UI commands links.editOriginal and links.reveal open the linked file in its app or show it in its folder (desktop). Edit Original runs only from the app's window, except for a placed document, which opens in a tab.

{"name":"run_command","arguments":{"command":"links.list","params":{"show":"missing"}}}
{"name":"run_command","arguments":{"command":"links.placementOptions","params":{"ids":[12],"preserve":"fit","align":"top"}}}
{"name":"run_command","arguments":{"command":"links.unembed","params":{"id":14,"path":"/tmp/logo.png"}}}

Package and Document Info

file.package {folder?, name?, copyLinks?, linksFolder?, relink?, copyFonts?, report?} copies a saved document (an unsaved one is an error) into folder/name (default name <document> Folder): <document>.vectorcraft, its linked files in Links/ (relinked: the packaged document points at the copies; the open one doesn't change), the fonts its type uses in Fonts/ (fonts whose licence doesn't allow embedding are listed in skippedFonts instead) and <document> Report.txt. A placed .vectorcraft document is packaged with its own linked files and fonts, relinked to the copies (each file copied once); one that can't be read is copied as it is, and a cycle of placed documents or one past the nesting limit stops there, each with a line in warnings and the report. A folder/name where a packaged file would replace the document or one of its linked files (the document's own folder, say) is an error, and nothing is written. Every option defaults to true. Without folder (the web, or an agent that wants the bytes) the result carries the same files as a zip ({name: "<name>.zip", dataBase64}, entries under <name>/).

document.info {selectionOnly?, category?, format?: "text"} adds sections ([{id, title, rows: [[label, value]]}]: document, objects, graphicStyles, brushes, spotColors, patterns, gradients, symbols, fonts, fontDetails, linkedImages, embeddedImages), brushNames (the Brushes panel's brushes; with selectionOnly, the brushes the selected objects' strokes use) and pathLength (pt: the total length of the paths, compound paths' pieces included and guides left out, which objects lists as Path Length in the document's units, say for a cut line or an LED strip along the letters); format: "text" returns the plain-text report Document Info › Save… (docInfo.save {path?}) and the package report write.

{"name":"run_command","arguments":{"command":"file.package","params":{"folder":"/tmp/handoff","copyFonts":false}}}
{"name":"run_command","arguments":{"command":"document.info","params":{"format":"text","category":"fontDetails"}}}

Slices

Object → Slice cuts the art into the pieces web output saves. User slices are rectangles of their own (ids from the object id counter); object.slice.make turns the selected objects into object slices, whose slice follows the object's bounds; auto slices fill what no other slice covers (the artboards with Clip to Artboard on, the default, else the art and the slices). slice.list returns every slice as laid out, numbered left to right and top to bottom: {number, id (null for auto slices), source: user|object|auto, name, x, y, width, height, options, selected}, plus clipToArtboard, hidden and locked.

The Object → Slice commands act on the selected slices: the ones select.object.slices (Select → Object → Slices) selects, and the selected objects that are object slices. object.slice.release (an object slice leaves its object, a user slice becomes an unpainted rectangle), object.slice.fromGuides (the grid the ruler guides cut the artboards into), object.slice.fromSelection, object.slice.duplicate {dx?, dy?}, object.slice.combine, object.slice.divide {rows? | rowHeight?, columns? | columnWidth?}, object.slice.deleteAll, object.slice.options {kind?: image|noImage|htmlText, name?, url?, target?, message?, alt?, text?, background?: ""|matte|#rrggbb, hAlign?, vAlign?} and object.slice.clipToArtboard {on?} are each one undo step. view.slices.hide {hidden?} and view.slices.lock {locked?} are session view state; the canvas draws the slices in the sliceLineColor preference, numbered while showSliceNumbers is on.

{"name":"run_command","arguments":{"command":"object.slice.make","params":{}}}
{"name":"run_command","arguments":{"command":"object.slice.options","params":{"name":"logo","url":"https://example.com","alt":"Logo"}}}
{"name":"run_command","arguments":{"command":"slice.list","params":{}}}

The Slice tool (slice, Shift+K) and the Slice Selection tool (sliceSelection) reduce to commands: object.slice.create {x, y, width, height} (the Slice tool's drag; it snaps to ruler guides), object.slice.select {slices?, toggle?}, object.slice.move {slices?, dx, dy} (an object slice moves its object), object.slice.setRect {id, x, y, width, height} (user slices) and object.slice.delete {slices?}. Slice ids are user slice ids and the ids of objects with an object slice.

{"name":"run_command","arguments":{"command":"object.slice.create","params":{"x":0,"y":0,"width":200,"height":80}}}
{"name":"run_command","arguments":{"command":"object.slice.move","params":{"dx":10,"dy":0}}}

Export for Screens

document.exportForScreens writes every chosen artboard (range, artboards; default all) in every format row into folder, or returns the files as dataBase64 without one; zip: true packs them into one store-only .zip. A row is {format, scale?} where scale is a factor (2, "2x"), a pixel width ("100w" or width: 100), a height ("100h") or a resolution ("72ppi"); raster rows name their files @2x, @100w… unless suffix says otherwise. fullDocument writes one file per row instead (a PDF of every artboard, other formats the bounds of all art), includeBleed grows each artboard by the document's bleed, subfolders puts each row's files in a sub-folder (its size or format, or its own folder), preset: "mobile" or "density" (Android-style ldpi…xxxhdpi sub-folders) replaces the rows, and settings: {png: {…}, jpg: {…}, svg: {…}, pdf: {preset}} gives every row of a format its options. On every platform, /, \, :, <, >, ", |, ?, * and control characters from a prefix, a suffix, a name pattern or an item's name become - in file and folder names, a folder name ends without periods or spaces, and a Windows device name (CON, PRN, AUX, NUL, COM1 to COM9, LPT1 to LPT9) that is the whole name or comes before its first period gets a - after it (CON-.png). The document remembers the last params: document.exportSettings. As in document.export, the reply's warnings lists the notes of every file exported (an image that couldn't be decoded, type written in the fallback font…), each note once; the export still writes or returns every file.

Batch exports (#1935):

• What: layers: [id…] (each layer's art alone) or selection: "one" | "each" instead of artboards (assets, fullDocument as before).
• Names: name is a pattern for every file: <filename> (the document), <name> (the artboard, asset, layer or object), <index>, #### (a counter zero-padded to the run of #, from counterStart; counterPer: "item" or "file"), YYYYMMDD, YYYY-MM-DD, <date>, <time>, <format>, <scale> and <suffix> (else the row's suffix ends the name; date sets the day the pattern uses). Different files it names alike get -2, -3….
• Folders: folders: "none" | "rows" | "format" | "item" | "itemFormat" (all in one, a folder per row, per extension, per item, or item then extension).
• Formats: rows may also be eps, ai and vectorcraft (a document of the item alone). A PDF row with marks ({trim, registration, colorBars, pageInfo}) and a suffix such as -print sits next to a plain PDF row.
• Presets: document.exportPresets.list (built-in: Logo Package, Print-Ready PDF), .save {name, settings} (kept with the preferences) and .delete {name}; exportPreset: "<name>" uses one under the call's own params.

{"name":"run_command","arguments":{"command":"document.exportForScreens","params":{"range":"1-3","name":"COMPANY-Logo-<filename>-####-YYYYMMDD","folders":"format","formats":[{"format":"svg"},{"format":"png"},{"format":"pdf"},{"format":"pdf","suffix":"-print","marks":{"trim":true,"registration":true,"colorBars":true,"pageInfo":true}}]}}}
{"name":"run_command","arguments":{"command":"document.exportForScreens","params":{"exportPreset":"Logo Package","layers":[2,9]}}}
{"name":"run_command","arguments":{"command":"document.exportForScreens","params":{"preset":"density","zip":true}}}
{"name":"run_command","arguments":{"command":"document.exportForScreens","params":{"range":"1-2","formats":[{"format":"png","scale":"512w"},{"format":"jpg","quality":80,"scale":"2x"},{"format":"svg"}],"settings":{"png":{"background":"white"}}}}}
{"name":"run_command","arguments":{"command":"document.exportSettings","params":{}}}

EPS export

document.exportEps (also document.export / export with format eps) writes Encapsulated PostScript: a DSC header (%!PS-Adobe-3.0 EPSF-3.0, %%BoundingBox, %%HiResBoundingBox, %%LanguageLevel, spot colours as %%DocumentCustomColors), then the page. Without useArtboards the file covers the visible art, its bounding box y up from the first artboard's bottom-left corner; useArtboards: true (or range, artboards, artboard) writes one file per artboard, {stem}_{artboard}.eps, each bounded by its artboard. level: 3 (default) writes gradients as smooth shadings and masks transparent image pixels out; level: 2 writes gradients as bands of colour (as does compatibleGradients) and transparent pixels white. Type is always glyph outlines, so no fonts are needed (embedFonts is accepted); spot colours are Separation colour spaces, and overprinting fills and strokes overprint unless overprints: "discard". RGB documents are written in CMYK unless cmykPostScript: false.

PostScript has no transparency: it is flattened first with flattenerPreset (high, medium (default), low or a saved preset, see flattener.presets.list) and flattener: {…} options over it (those of object.flattenTransparency); a warning says so. previewFormat adds a TIFF preview behind a binary header (tiffColor (default), transparent unless transparentPreview: false, or 1-bit tiffBw; none writes plain PostScript), thumbnails a PNG thumbnail in comments, and the native document always rides along in the trailer's comments (with linked images whole when includeLinkedFiles), so a later EPS import can reopen the file as it was. Export for Screens doesn't write EPS.

{"name":"run_command","arguments":{"command":"document.exportEps","params":{"path":"/tmp/logo.eps","level":2,"previewFormat":"tiffBw","flattenerPreset":"high"}}}
{"name":"export","arguments":{"path":"/tmp/art.eps","options":{"useArtboards":true,"cmykPostScript":false}}}

Print

File → Print lays the job out as a print-ready PDF, in which each page is a sheet of the chosen paper. The desktop app sends the PDF to the system's print service, and the web build opens the browser's print dialog on it. With PDF File picked as the printer, File → Print saves the PDF instead (see Printing). Headless, file.print {settings?, path?} → {pages, path | dataBase64, bytes, warnings}; print.preview {settings?} answers what it would print without writing it: {pages (copies included), sheets: [{artboard, tile?, ink?, width, height, orientation, scale}] (one copy, in order), tiles: [{artboard, columns, rows, printed, tiles}] (document space), inks: [{name, spot, print, frequency, angle}], warnings, settings}. print.setup {settings?} saves the settings with the document (one undo step, kept in the native file as print_setup); both other commands start from the saved settings, or from a print preset's with preset (see Print presets), with settings over them (null keeps a value). list_commands with filter print.setup documents every field:

• General: copies (1–999), collate, reverse; artboards: all|range|ignore with range: "1-3, 5" (ignore: all the art as one page), skipBlank; media: letter|legal|tabloid|a3|a4|a5|b4|b5|custom (width, height in pt), orientation, autoRotate (on by default: the paper turns to each artboard), transverse; printLayers: visiblePrintable|visible|all (template layers never print); placement: {origin, x, y} places the printed area (artboard, bleed and marks) on the imageable area (the paper inside margin); scaling: none|fit|custom|tileFull| tileImageable with scale: {width, height} (%), overlap and tileRange (tiles numbered across then down).
• marks and bleed as document.exportPdf takes them (the bleed is the document's unless useDocument: false), drawn around the artboard at the paper's scale; the page information adds the tile and the ink with its screen.
• output: {mode: composite|separations, emulsion: up|down, image: positive|negative, spotsToProcess, inks: [{name, print, frequency, angle}]}: separations print one page per ink in the grey of its coverage (images too; overprinting fills and strokes don't knock out), Registration on every plate; emulsion down mirrors the page and a negative inverts it.
• graphics: {autoFlatness, flatness, fonts} and color: {intent, preserveNumbers} (the intent colours separate with; without preserveNumbers CMYK colours are separated again).

Halftone screens and a fixed flatness are not written to the PDF (the output device's apply): the warnings say so, as they do for art larger than the imageable area. Type in a font family that isn't available prints in the fallback font, and the warnings of file.print and print.preview name those families. They count only type that prints: visible type that fits its frame or path, on the layers printLayers picks, on the paper of the artboards or tiles that print, and the type in the symbols, patterns, scatter brushes, opacity masks and placed documents of the art that prints. The warnings also give the ids of the area type and type on a path that print and whose text doesn't fit (objects of the document only, not type in the art of symbols, patterns, brushes, opacity masks or placed documents; see Fitting area type).

{"name":"run_command","arguments":{"command":"print.setup","params":{"settings":{"media":"a4","scaling":"fit","marks":{"trim":true}}}}}
{"name":"run_command","arguments":{"command":"print.preview","params":{"settings":{"output":{"mode":"separations"}}}}}
{"name":"run_command","arguments":{"command":"file.print","params":{"path":"/tmp/job.pdf","settings":{"copies":2}}}}

Print to a PostScript file: file.print {format: "postscript"} (the default for a .ps path) writes the same job as PostScript, level: 3 (default) or 2, answering format: "postscript" ("pdf" otherwise; in the app the job is saved, not sent to a printer). The file follows the DSC: %%Pages is the page count (copies included), each %%Page sets its paper size, separations name their ink (%%PlateColor) and set its halftone screen (frequency, angle, a round dot), a fixed flatness is written (setflat), negatives invert the transfer, and the marks print in Registration. PostScript has no transparency: it is flattened first with flattenerPreset (medium by default; high, low or a saved one), with a warning.

{"name":"run_command","arguments":{"command":"file.print","params":{"path":"/tmp/job.ps","level":2,"settings":{"output":{"mode":"separations"}}}}}

Data Recovery

Modified documents get recovery copies (file.recovery.save; the app runs it every autosaveInterval minutes while autosaveRecovery is on, skipping documents of more than 20,000 objects while recoveryOffForComplex is on). A copy goes when its document is saved, reverted or closed, so copies still there after the app quits were left by a crash. The copies live in the recoveryFolder preference's folder (default: Data Recovery beside the app's preferences; none when the app runs with VECTORCRAFT_NO_PREFS) or, on the web, in browser storage.

Each running app (each browser tab) keeps its copies in an area of its own (<area>/<name>): a sub-folder whose .lock file it keeps locked while it runs, or on the web an area with a heartbeat it refreshes every minute and a Web Lock the browser holds until the tab is gone (where the browser has Web Locks: secure pages). Only areas nobody holds are offered: their lock is free, or no tab holds their Web Lock and their heartbeat is older than three intervals (at least three minutes). A background tab whose timers are paused therefore keeps its copies. Without Web Locks another tab can take such a tab for gone; when it resumes, its next heartbeat or recovery save writes its missing copies again. Several apps running at once (agents' instances included) never see each other's copies as crash leftovers, and an area being restored or discarded is held, so two apps launched together never both take it.

file.recovery.list → {copies: [{file, title, path, format, saved, open, running}], location} (open: the copy of a document open here; running: kept by another VectorCraft that is running). file.recovery.restore {file?} opens copies left behind (default: all) as "<name> [Recovered]": modified, and Save asks where to save them (suggesting their original file); the copy moves into this app's area. file.recovery.discard {file?} deletes them. Without a folder or browser storage the commands are disabled and say how to set one.

{"name":"run_command","arguments":{"command":"prefs.set","params":{"key":"recoveryFolder","value":"/tmp/vc-recovery"}}}
{"name":"run_command","arguments":{"command":"file.recovery.save","params":{}}}
{"name":"run_command","arguments":{"command":"file.recovery.restore","params":{"file":"1759650000-1/Poster-1"}}}

Native save options

document.save / file.saveAs / file.saveCopy to a native (.vectorcraft, .vctemplate) or .ai file take, flat or in options (file.formatOptions {format} lists them with their values):

• separateArtboards: true also writes each artboard of range ("1-3, 5", default "all") to <name>-<artboard>.<ext> beside the file: that artboard and the art touching it. The result's files lists every file written, the master file first (without a path: files: [{name, dataBase64}]). 3 artboards give 4 files.
• includeLinked: true keeps linked images' own pixels in the file (they stay linked), not just the previews.
• embedProfiles (default true) carries the ICC profiles the document is tagged with that were loaded from files; opening the file installs them where they are missing (a warning names one that can't be used).
• pdfCompatible: native files (default false) also carry a PDF of every artboard (pdf in the file); .ai files (default true) with false write blank pages around the native document (smaller; other apps show empty pages).
• .ai only: compress (default true) is the PDF option compression.compressText.

File → Save As asks for these in the save options dialog (VectorCraft Options) after the save panel, and Save a Copy does too for a .ai file; Save, Save a Copy to a native file and Save as Template reuse the options as last saved (or Use Compression).

{"name":"run_command","arguments":{"command":"file.saveAs","params":{"path":"/tmp/set.vectorcraft","separateArtboards":true,"range":"2-3","includeLinked":true}}}
{"name":"run_command","arguments":{"command":"file.saveAs","params":{"path":"/tmp/set.ai","pdfCompatible":false}}}

EMF and WMF

document.export (and export) with format emf or wmf writes a Windows metafile of the first artboard (or artboard); useArtboards: true writes one file per chosen artboard ({stem}-{artboard}.emf), false the bounds of the visible art. EMF keeps Bézier curves, solid fills (brushes), strokes as geometric pens with their width, caps, joins, miter limit and dashes, clipping masks (path clips), images with their transparency (AlphaBlend) and type as outlines; gradients become images clipped to their shape, pattern fills their tiles clipped to the shape, and raster effects images rendered at the document's raster effects resolution. The header's frame is the artboard in 0.01 mm. WMF flattens curves into polygons in 16-bit units behind a placeable header (1440 units an inch, fewer for pictures over about 22 inches); it has no clipping, transparency or gradients, so clipped art is written whole, images over white and gradients and patterns as one colour. What a format leaves out comes back in warnings.

document.open and file.place read .emf and .wmf (placeable or not) into one artboard, the picture's frame, and one layer: paths with their fills and strokes (pens become strokes with caps, joins and dashes), clipping groups, images and point type in the font the file names. Records VectorCraft doesn't read (EMF+ drawing among them) are skipped with one warning. clipboard.importEmf {dataBase64, center?} loads an EMF or WMF picture into the clipboard for edit.paste; the app pastes image/emf from the system clipboard ahead of text and bitmaps when the platform's clipboard service offers it (the Windows desktop service doesn't read metafiles yet).

{"name":"export","arguments":{"path":"/tmp/logo.emf"}}
{"name":"run_command","arguments":{"command":"document.export","params":{"format":"wmf","useArtboards":true,"path":"/tmp/icons.wmf"}}}
{"name":"run_command","arguments":{"command":"file.place","params":{"path":"/tmp/chart.emf","at":[300,200]}}}

TIFF, BMP and Targa export

document.export (and export) writes them like the other raster formats: one artboard (or useArtboards, one file each), at ppi, over background, with antiAlias. TIFF takes colorModel (rgb keeps transparency as an unassociated alpha channel, cmyk writes ink amounts in the working CMYK space with CMYK colours keeping their inks, gray is flattened on white), lzw (default true), byteOrder (little, the II files PCs write, or big, MM) and embedIcc (sRGB, the working CMYK profile or grey). BMP takes depth (1 is black and white; 4 and 8 use a palette chosen by reduction and dither; 16, 24, or 32, which keeps transparency), colorModel (rgb or gray), fileFormat (windows or os2: 1, 4, 8 or 24 bits only), rle (RLE4/RLE8 for 4- and 8-bit Windows bitmaps) and flipRows (rows top-down, a negative height; not with rle or os2); impossible combinations are refused. Targa takes depth: 16 (a one-bit alpha), 24 (default) or 32 (with alpha). Formats and depths without alpha are flattened on white. TIFF, BMP and Targa files open again (a 16-bit Targa file with its one-bit alpha); document.formats lists every option.

{"name":"export","arguments":{"path":"/tmp/print.tif","options":{"ppi":300,"colorModel":"cmyk","byteOrder":"big"}}}
{"name":"run_command","arguments":{"command":"document.export","params":{"format":"bmp","depth":8,"rle":true}}}
{"name":"export","arguments":{"path":"/tmp/sprite.tga","options":{"depth":32}}}

Save for Web

File → Export → Save for Web (Legacy) writes optimised web images: document.exportForWeb {format: gif|jpg|png8|png24, …} renders the artboard (clipToArtboard, the default; artboard picks one) or the visible art and the slices, at width, height or percent, and encodes it with the format's settings: for GIF and PNG-8 reduction, colors, dither, ditherAmount, transparency, matte, interlaced, webSnap, lossy (GIF) and colorTable {locked, transparent, webShift, sort}; for JPEG quality, progressive, optimized, embedProfile and matte; for PNG-24 transparency, matte and interlaced. metadata (none, copyright, contact, all) writes File Info as PNG text or a GIF/JPEG comment; convertToSrgb marks PNGs as sRGB. <stem> is the name of path without its extension, or without path the document's title with the characters Windows does not allow in a file name as - and a - after a Windows device name. With slices (slice.list) each image slice becomes images/<slice name>.<ext>, with a - after a Windows device name such as aux (slices: selected writes the selected ones, none one image); output: html adds <stem>.html placing them, with each slice's URL, target, alt text and status message, and No Image / HTML Text cells. Without path the files come back as {name, dataBase64}. As in the raster exports, warnings says when an image couldn't be decoded and is missing from the image; the files are written all the same.

document.exportForWeb.preview returns what Save writes for the same settings without writing it: bytes (the exact file size), seconds to download at kbps (default 56.6), the size, and for palette formats the colour table (colors: [{color, source, transparent, locked, webShifted, webSafe}]). The colour table's lists name colours by source (the colour the reduction made), so a colour keeps its edits when it is web-shifted or snapped. slice: n previews one slice; image: true adds the file; warnings holds the export's warnings.

Presets: webExport.presets.list (built-in ones, then saved ones), webExport.presets.save {name, newName?, …settings}, webExport.presets.delete {name}; any name works as preset in the other commands (its settings, then the keys given). webExport.settings {…} remembers the settings the dialog opens on ({} reads them, reset: true restores the defaults); presets and remembered settings are preferences.

{"name":"run_command","arguments":{"command":"document.exportForWeb.preview","params":{"format":"gif","colors":32,"dither":"none"}}}
{"name":"run_command","arguments":{"command":"document.exportForWeb","params":{"path":"/tmp/web/page.html","format":"png8","output":"html"}}}
{"name":"run_command","arguments":{"command":"webExport.presets.save","params":{"name":"Banner","format":"jpg","quality":70}}}

Asset Export

Assets are pieces of art collected for export on their own (Window › Asset Export, Object › Collect for Export, the Layers panel ≡ menu and row menu, File › Export Selection…). assets.add {ids?, multiple?} collects the selected objects (or ids): one asset per object, or with multiple: false one asset of them all, named after the object or Asset 1, Asset 2…; a layer id collects the art under it (sublayers included). Art already collected keeps its asset. Assets name their objects, so they follow edits: moving the art moves the crop, and deleting it removes it from its asset (an asset left with nothing goes). assets.list gives each asset's id, name, objects and crop (bounds); assets.rename {asset, name} and assets.remove {assets} are undo steps.

assets.export {assets?, folder?, zip?} writes each asset's art alone, cropped to it and named after it, in every format row of the document's export settings, the ones Export for Screens remembers (document.exportSettings); any formats, preset, settings, prefix or subfolders given win for that export. assets.settings.set changes those shared settings (not an undo step). document.exportForScreens {assets: [id…]} does the same with its own params (the dialog's Assets tab). An asset's crop, and the crop of export with selection: true, is the bounds the renderer measures for what the art draws, its effects included.

{"name":"run_command","arguments":{"command":"assets.add","params":{"multiple":true}}}
{"name":"run_command","arguments":{"command":"assets.add","params":{"ids":[12],"multiple":true}}}
{"name":"run_command","arguments":{"command":"assets.settings.set","params":{"formats":[{"format":"png","scale":"1x"},{"format":"png","scale":"2x"},{"format":"svg"}]}}}
{"name":"run_command","arguments":{"command":"assets.export","params":{"zip":true}}}

A layer id in ids (12 above) is the same as Collect for Export on that Layers row: every object under the layer becomes an asset (or one asset with multiple: false).

CSS Properties

css.selection {ids?, units?, position?, dimensions?, unnamed?, rasterize?} returns the CSS web pages style the selected objects (or ids) with, one rule per object, as Window › CSS Properties shows it: a rectangle's or ellipse's fill as background-color or a linear-gradient()/radial-gradient(), its stroke as border, its corners as border-radius; type's font, color, spacing and alignment; opacity, blend mode, shadows and glows (box-shadow, text-shadow) and blur (filter). Layers and plain groups stand for what they hold. Properties SVG export writes too (font, spacing, opacity, blend mode, colours, gradient stops) read exactly as in its style sheets. Each rule in rules says when CSS can't describe the art exactly (unsupported: other shapes, images, patterns, live effects…); it is then written as its box, or with rasterize: true as background-image: url(<class>.png). units is px (1 px per point, as in SVG export), pt, mm, cm or in; unnamed: false leaves out unnamed objects (counted in skipped). css.generate does the same for the whole document, and css.export {path?, scope?: selection|all} writes the .css file with the PNGs of rasterized art next to it (no path: data and images as base64). Its warnings add the PNGs' notes to the CSS's, each once (an image that couldn't be decoded, type written in the fallback font…). In the app, css.copy copies the CSS and css.exportFile writes it through a save panel (the web downloads it) and shows the first note in the status bar.

{"name":"run_command","arguments":{"command":"css.selection","params":{"units":"px","position":true}}}
{"name":"run_command","arguments":{"command":"css.export","params":{"path":"/tmp/site/styles.css","scope":"all","rasterize":true}}}

PSD export

document.export (and export) writes layered bitmaps (.psd, 8 bits per channel) like the other raster formats: one artboard (or useArtboards, one file each), at ppi, with antiAlias, in colorModel rgb, cmyk (ink amounts in the working CMYK space) or gray, with embedIcc. With layers (default true) each top-level layer is a pixel layer of its own, drawn alone and cropped to what it paints, with its name, opacity and blend mode; a background colour becomes a Background layer under them, and the merged image (what readers that skip layers show) is the flat render. maxEditability: true turns layers and sublayers into groups and every object into a layer named as the Layers panel names it (text objects by their text); layers with a clipping mask, an opacity mask, an appearance of their own or knockout stay one pixel layer. Hidden layers and objects are left out unless hiddenLayers: true writes them as hidden layers. layers: false writes one flat image, on white where nothing is drawn. Files are at most 30000 pixels a side.

Photoshop documents (.psd, and .psb, the large document format) open and place as their merged image, the flattened picture every Photoshop file carries (layers aren't read): Bitmap, Grayscale, Duotone, Indexed, RGB, CMYK (converted with the colour settings' CMYK) and Lab, at 1, 8, 16 and 32 bits per channel, sized by their resolution. A document with transparency keeps it. file.place {path} links the file like any image (links.update reads it again when it changes); Multichannel files are refused.

{"name":"export","arguments":{"path":"/tmp/poster.psd","options":{"ppi":300,"maxEditability":true}}}
{"name":"run_command","arguments":{"command":"document.export","params":{"format":"psd","layers":false,"colorModel":"cmyk"}}}

Printing

print.setup {settings?} keeps print settings with the document (one undo step; without settings it answers the current ones): the copies, artboards, paper, orientation, placement, scaling and tiling, marks and bleed, composite or separations output, graphics and colour management options of File › Print. print.preview {settings?} lays the job out without printing: pages, then one copy's sheets, each with its page size, scale, the transform from the document onto the page (pt, y down), the document area it prints and its trim box on the page, plus tiles, the inks of a separation and warnings. file.print {settings?, path?} writes the job as a print-ready PDF (one page per sheet, tile, ink and copy); headless it answers dataBase64. In the app, file.print with no params opens the Print dialog, and with params it prints through the system (printer: a name from print.printers, default the system's default printer) or, with toFile, a path or no printing available, saves the PDF.

{"name":"run_command","arguments":{"command":"print.setup","params":{"settings":{"media":"a4","scaling":"fit","marks":{"trim":true}}}}}
{"name":"run_command","arguments":{"command":"print.preview","params":{"settings":{"copies":2,"output":{"mode":"separations"}}}}}
{"name":"run_command","arguments":{"command":"file.print","params":{"path":"/tmp/job.pdf"}}}

Print presets

print.presets.list lists [Default] (the default print settings, builtIn: true, protected) and the saved presets, each with its settings and how they differ from [Default] (changed). print.presets.save {name?, newName?, preset?, settings?} creates or changes a saved preset (starting from its own settings, or from preset, with settings over them) and print.presets.delete {name} deletes one; [Default] can be neither changed nor deleted. print.presets.export {names?, path?} writes them as a .vcprintpresets JSON file (without path it returns data), and print.presets.import {path? | data? | dataBase64?, replace?} adds a file's presets (a name in use gets a number unless replace). Saved presets live with the preferences. print.setup, print.preview and file.print take preset, the name of [Default] or of a saved preset in any case: the preset's settings are used instead of the document's, with settings over them, and print.setup {preset} saves them with the document. An unknown name is refused with the list of presets.

{"name":"run_command","arguments":{"command":"print.presets.save","params":{"name":"Posters","settings":{"scaling":"tileImageable","overlap":18,"marks":{"trim":true}}}}}
{"name":"run_command","arguments":{"command":"file.print","params":{"preset":"Posters","path":"/tmp/posters.pdf"}}}

EPS and PostScript import

document.open and file.place read .eps, .epsf and .ps files, and PostScript by any name (an .ai saved without PDF compatibility opens as artwork; an .ait one as a new untitled document). A .ps file reports the format ps, which document.export doesn't write (file.print writes PostScript). An EPS file VectorCraft wrote restores the document it carries (restored: true, as Preserve Editing does for PDF and SVG); when that document can't be read the PostScript is read instead and the first warning says why.

Other files are run through a small PostScript interpreter (Level 3, first page only): paths, fills and strokes with their width, caps, joins, miter limit and dashes (a fill and a stroke of the same path become one object), grey, RGB, CMYK, indexed and spot colours (a Separation ink becomes a spot swatch, painted at its tint), clips (clipping groups; clipsave/cliprestore), gsave/grestore, save/restore, transforms, procedures with bind def, loops, dictionaries (the standard ones are values in systemdict; dictstack, internaldict), arrays and strings whose intervals share storage, resources (categories made from Generic, resourceforall), executable filters (cvx exec runs their data; flushfile skips it), axial and radial shadings and shading patterns (gradients), images and image masks (data in the file through ASCII85, hex, run-length, Flate, LZW or DCT filters, or from procedures), and type as point type in the font the file names (embedded font programs are skipped). In a file in Illustrator's own format (Illustrator 3–8 .ai and their EPS, written with the prolog that defines its operators) the groups it writes (u … U) come in as groups, nested as they were; clips and groups nest at most 128 deep. The artboard is the %%HiResBoundingBox (else %%BoundingBox; a letter page without one). A program the interpreter can't run (an operator it doesn't know, an error, a runaway loop) or that draws nothing comes in as its TIFF preview (palette previews with an alpha channel too) with a warning; without a preview, the art drawn up to the error is kept with a warning, and a file with none is refused with a message saying why.

{"name":"run_command","arguments":{"command":"document.open","params":{"path":"/tmp/logo.eps"}}}
{"name":"run_command","arguments":{"command":"file.place","params":{"path":"/tmp/logo.eps","at":[300,300]}}}

PDF/X

document.exportPdf {standard: "pdfX1a" | "pdfX3" | "pdfX4"} (or the built-in PDF/X presets) writes print exchange files as their standards say. Every PDF/X file gets a /GTS_PDFX output intent (output.outputIntent; blank, the CMYK profile in effect: the destination when converting to CMYK, else the working one, embedded; in PDF/X-1a and PDF/X-3 a name that isn't a profile needs output.registry, a registered printing condition named without a profile; PDF/X-4 always embeds one), a TrimBox on every page (the artboard), /Trapped True or False, /GTS_PDFXVersion, a title (Untitled when the document has none) and dates; type is outlined or embedded; editing data, passwords and, in PDF/X-1a and PDF/X-3, PDF layers are refused.

• pdfX1a (PDF/X-1a:2001, written as PDF 1.3): CMYK, grey and spot colours only. Every colour and image is converted to output.destination (blank: the output intent's CMYK profile, else the working CMYK profile), untagged; an RGB destination or output intent is refused. Transparency is flattened before writing.
• pdfX3 (PDF/X-3:2002, written as PDF 1.3): colours are tagged with ICC profiles; transparency is flattened.
• pdfX4 (PDF/X-4:2010, PDF 1.6 at most): transparency and layers are kept, colours are tagged, and the XMP metadata names the standard.

Flattening uses flattenerPreset (default High Resolution, see PDF 1.3 above) on a copy (the document is untouched; a warning says so), and images with see-through pixels count as transparency. The written file is checked against its standard: what it still breaks (transparency, RGB in PDF/X-1a, a font not embedded) fails the export instead of writing a file that only claims the standard.

{"name":"run_command","arguments":{"command":"document.exportPdf","params":{"path":"/tmp/press.pdf","preset":"PDF/X-1a:2001","output":{"outputIntent":"VectorCraft Generic CMYK (SWOP-like)","outputCondition":"Coated","trapped":false}}}}

Print tiling

View → Show Print Tiling (view.printTiling {on?}, per document, not undoable) draws the pages of the document's print settings on the canvas: each page's paper edge, its imageable area dashed, numbered (tiles across then down; tiles outside tileRange dimmer). The Print Tiling tool (printTiling) shows them too. print.tiling {settings?} answers what is drawn, from the layout print.preview and both writers (PDF and PostScript) use, so the canvas and the printed job agree: {tileOrigin, pages: [{artboard, number, printed, page, imageable}]} ([x0, y0, x1, y1] in document space; every tile when tiling, else each page once, not once per ink).

Dragging with the tool puts the top-left corner of the first page's imageable area where the pointer is (snapping to the artboard's edges): print.tiling.set {origin: [x, y], artboard?} saves it with the print settings as tileOrigin: {placed: true, x, y}, measured from the artboard's top-left corner (one undo step). It replaces the placement, and tiles start there (going on left and up as far as the art does). A double click on the canvas, or on the tool's button, runs print.tiling.set {reset: true}: the placement places the pages again. In the Print dialog the placement then shows as placed by the tool, with a Reset button; dragging the preview moves the origin.

{"name":"run_command","arguments":{"command":"view.printTiling","params":{"on":true}}}
{"name":"run_command","arguments":{"command":"print.tiling.set","params":{"origin":[-36,-36]}}}
{"name":"run_command","arguments":{"command":"print.tiling","params":{"settings":{"scaling":"tileImageable","scale":{"width":300,"height":300}}}}}

Print advanced options

The Print dialog's Advanced section and the printer profile are print settings too (print.setup, print.preview, file.print): advanced: {printAsBitmap, overprints: preserve|discard|simulate, flattenerPreset} and color: {profile}. In composite output, preserve writes overprinting fills and strokes with an overprinting graphics state (/OP, as PDF export does), discard makes overprinting fills and strokes knock out and simulate prints them as Overprint Preview shows them (darkened); either way the file has no overprint left. Separations always keep overprints. flattenerPreset (High Resolution, Medium Resolution, Low Resolution or a saved preset, flattener.presets.list) flattens transparency before printing, as Object → Flatten Transparency does; empty, a PDF keeps it live (PostScript, which has none, uses file.print's flattenerPreset, else this one, else medium). printAsBitmap prints each composite page as one image of the art at the document's raster effects resolution. color.profile names an RGB or CMYK profile: composite PDF colours are converted to it with color.intent (CMYK colours keep their numbers with preserveNumbers), so a CMYK profile writes device CMYK; separations separate with a CMYK one. Settings saved before these options load with their defaults.

{"name":"run_command","arguments":{"command":"file.print","params":{"path":"/tmp/proof.pdf","settings":{"advanced":{"overprints":"simulate","flattenerPreset":"High Resolution"},"color":{"profile":"VectorCraft Generic CMYK (SWOP-like)","intent":"perceptual"}}}}}

Rotated bounding boxes

Every object keeps the angle of its own axes (counter-clockwise degrees; document.inspect reports it as a node's rotation, and the selection's shared angle as selectionRotation, 0 when the selected objects differ). Transforms turn it with the object while its path geometry is still baked, so after a rotation the bounding box and its handles stay square to it, as in the reference app. object.rotate {angle, absolute: true} turns the selection to an angle (the Transform and Properties panels' Rotate field), and object.setBounds measures width, height and the reference point along the turned box (x, y stay page coordinates). Objects turned together share the box. A new group starts square to the page. object.resetBoundingBox squares the box again without moving the art.

{"name":"run_command","arguments":{"command":"object.rotate","params":{"angle":45,"absolute":true}}}

Select All while editing type

While the Type tool edits text (a click into type, pointer_gesture with "tool":"type"), select.all (Cmd+A, press_key {key: "A", mods: {cmd: true}}) selects all of that text instead of the art, as in the reference app, and returns {editing, start, end} (the text's id and the selected byte range, which text.getRange and text.setRangeStyle take); the art selection stays as it is. Without text being edited it selects every object and returns {count}. Text in threaded frames is selected one frame at a time.

{"name":"press_key","arguments":{"key":"A","mods":{"cmd":true}}}

Empty point type

Point type the Type tool places with a click and leaves empty is discarded when editing ends: Escape, another tool, a click elsewhere, switching documents, or a command that takes the text out of the selection (select.none, a select.set of other objects). The steps since the click go with it (text.discardEmpty {id}), so the document and its undo history are as if the click never happened. Commands on the edited text (the Character panel's text.setRangeStyle {id}) keep it in editing. Area type dragged as a frame and type in or on a path keep their frame (Object → Path → Clean Up removes empty text).

{"name":"run_command","arguments":{"command":"text.discardEmpty","params":{"id":42}}}

Resizing area type

As in the reference app, dragging a bounding-box handle of area type resizes its type area instead of scaling the type: the Selection tool sends object.transform {matrix, typeAreas: true}, which reshapes the frame of each area type object it transforms (other objects, type inside groups and point type transform as usual) and reflows the text at its size, through its thread too. text.reshapeArea {id, anchors: [[subpath, anchor]…], dx, dy} moves frame anchors with their handles (Direct Selection dragging a corner, or an edge's two ends), and text.areaOptions {width?, height?} sizes the area from its top-left corner (the query returns both). Each is one undo step. Object › Transform › Scale, the Scale tool and the Transform panel still scale the type.

{"name":"run_command","arguments":{"command":"text.reshapeArea","params":{"id":42,"anchors":[[0,2]],"dx":40,"dy":60}}}

Vertical alignment

text.areaOptions {verticalAlign: top|center|bottom|justify} (Area Type Options › Align) places the lines of each row/column cell on its own: center centres the block of lines in the cell, bottom puts the last line's descent on the cell's bottom (inside the inset), justify keeps the first line at the top, moves the last one to the bottom and shares the space left over equally between the lines (there is no paragraph spacing limit: every line gap grows by the same amount). A cell with a single line justifies to the top, and a full or overflowing cell only moves by the less-than-a-line of space it has left. In rectangular frames without text wrap the lines just move; in other frames (and around wrap objects) the text flows again starting lower until it settles, never losing text that fit top-aligned. Vertical type aligns along its block axis (bottom is the frame's left edge). Exports lay type out again, so SVG, PDF and EPS show the aligned text.

{"name":"run_command","arguments":{"command":"text.areaOptions","params":{"verticalAlign":"center"}}}

Fitting area type: Auto Size, Shrink Text to Fit, overflow

text.areaOptions (and text.create's area, text.createInPath in area mode) take a fit:

• none (default): the frame keeps its size; text that doesn't fit overflows (the red "+").
• autoHeight: Auto Size. After every edit (typing, text.editRange, character and paragraph changes, a new width) the frame's bottom moves to just below the last line plus the inset, in the same undo step as the edit. In several columns the frame gets the least height (to 0.01 pt) at which the columns hold the text. It applies to rectangular frames of horizontal type in one row; height is ignored while it is on, and setting the height by hand (a handle drag, text.reshapeArea moving the bottom) turns it off, as does threading the frame. The autoSizeAreaType preference ("Auto Size New Area Type") gives new area type autoHeight.
• shrinkText (beyond the reference app), with fitMinPercent (10–100, default 50): when the text overflows, every run's size, leading (explicit leading; auto leading follows the size) and baseline shift are scaled by the largest factor down to fitMinPercent % that makes it fit. Paragraph spacing and indents stay. The stored sizes don't change: the scaling happens at layout time, so rendering and every export see it. The factor is found by bisection over the first run's size in steps of 0.1 pt (its scaled size is a whole number of tenths of a point, so the result is deterministic); at fitMinPercent the text may still overflow.

The file stores "fit": "autoHeight" or "fit": {"shrinkText": {"minPercent": 40}} in the object's area; the command also accepts that object form. The query (and every reply) of text.areaOptions reports fit as its id, fitMinPercent, overflow (the text doesn't fit its frame) and fitScale (Shrink Text's factor, 1 unshrunk). document.inspect (and document.node {summary: true}) reports overflow for area type and type on a path (text past the end of the path), and fit and fitScale for area type.

Text that doesn't fit is drawn up to the last line that does, so a one-line headline in a frame lower than its line draws nothing. The commands that make type or change what it holds say so in their own reply: text.create, text.createInPath, text.setText and text.setStyle add overflow: [id…] (the area type and type on a path among the objects they made or changed whose text doesn't fit), and the add_text tool adds overflow: true. The key is left out when everything fits. text.create and text.createInPath also return bounds: {x, y, width, height}, the new type's box as document.inspect reports it (after justify), so a script can place the type without a helper object (#2413). The warnings of file.print and print.preview give the ids of the objects that print and whose text doesn't fit: the first 20 ids, then how many more.

In PDF, EPS, SVG, EMF, WMF and the raster formats, the warnings of document.export, document.exportPdf and document.exportEps give the ids in the same way for the visible objects that the export draws and whose text doesn't fit. The export draws the objects on the pages or artboards it writes; an EPS of the visible art and an SVG without chosen artboards draw all of them. A PDF leaves out the layers that don't print unless includeNonPrinting or createLayers is set. A .ai file keeps the whole text in its editing data and gives no such warning.

{"name":"run_command","arguments":{"command":"text.areaOptions","params":{"fit":"shrinkText","fitMinPercent":60}}}

Threading text from the out port

As in the reference app, a click with the Selection tool on the red "+" out port of selected area type whose text overflows loads the pointer with that text (the port sits on the frame's right side, just above the bottom-right corner, so the corner handle stays free). The next click threads the text into the closed path or area type under the pointer, which becomes the next frame; a click on empty canvas makes a new frame of the first one's shape and size with its top-left corner at the click, and a drag draws a rectangle for the new frame. Escape or another tool drops the loaded text. Each placement is one undo step (text.thread.add), which agents can run directly:

text.thread.add {from, to? | rect?: [x, y, width, height] | at?: [x, y]} threads area type from into one more frame, put right after it in its thread (a new thread of the two when from has none): to, a closed path (converted to an empty frame) or area type that is not threaded yet; rect, a new rectangular frame (positive width and height); or at, a new frame of from's shape and size whose top-left is at at. Give exactly one of the three. New frames take from's type options, paint and first character and paragraph attributes, and go just above it. The story reflows through the thread, the two frames are selected, and it returns {id}, the frame threaded into. It refuses an id that doesn't exist, a from that isn't area type (type on a path doesn't thread), an open path, from itself, an object inside a compound path, a frame already in a thread, and coordinates that aren't finite numbers.

{"name":"run_command","arguments":{"command":"text.thread.add","params":{"from":42,"at":[300,120]}}}

Converting between point type and area type

With the Selection tool, a single selected point or area type object shows the type widget: a small circle beside the middle of its bounding box's right side, hollow on point type and filled on area type. Double-clicking it runs type.convertToAreaType (point type gets a frame around its text, which keeps its place and doesn't rewrap) or type.convertToPointType (soft line wraps become line breaks), the same commands as the Type menu's Convert To Area Type and Convert To Point Type. Either keeps the text and its styles and is one undo step; ids picks the objects (default: the selection).

{"name":"run_command","arguments":{"command":"type.convertToAreaType","params":{"ids":[42]}}}

Hyphenation options

text.setFormat sets the Paragraph panel menu's Hyphenation… options of the selected type (or ids, or the paragraphs a start / end range touches), which apply while hyphenate is on:

• wordsLongerThan (letters, 1 to 25, default 5): only words with more letters than this are hyphenated.
• afterFirst and beforeLast (letters, 1 to 15, default 3): the fewest letters a word keeps before and after the hyphen.
• hyphenLimit (lines, 0 to 25, default 0): the most consecutive lines that may end with a hyphen; 0 sets no limit. A line that ends at a soft hyphen counts, and a soft hyphen still breaks a line past the limit.
• hyphenationZone (pt, 0 to 1296, default 0): with the Single-line Composer, a line whose text before the next word already ends within this distance of the right edge breaks before that word instead of hyphenating it. The Every-line Composer does not use it.
• hyphenateCapitalized (default true): false leaves words that start with a capital letter whole. Words in capitals only are never hyphenated.

The letter and line counts are rounded, and every value is clamped to its range. Preferences › Hyphenation › Exceptions still decide the breaks of a listed word. The options are saved with the paragraph (hyphenation in para) only when they differ from the defaults, and paragraph styles carry them (paraStyle.new {attrs: {hyphenation: {words_longer_than: 8}}}). One undo step. ui.hyphenation opens the Hyphenation dialog on the values the Paragraph panel shows; OK sends the values that changed.

{"name":"run_command","arguments":{"command":"text.setFormat","params":{"hyphenate":true,"wordsLongerThan":8,"hyphenLimit":2}}}

Justification options

text.setFormat sets the Paragraph panel menu's Justification… options of the selected type (or ids, or the paragraphs a start / end range touches), in percent:

• wordSpacing (0 to 1000, default 80 / 100 / 133), letterSpacing (-100 to 500, a share of the font's word space, default 0 / 0 / 0) and glyphScaling (50 to 200, default 100 / 100 / 100), each [min, desired, max] or {min?, desired?, max?} (the values not given are kept). Desired applies to every line: word spaces are that share of the font's word space, each letter followed by another is followed by that share of a word space, and glyphs are that wide. A justified line may move them between Minimum and Maximum to fill its measure: word spacing first, then letter spacing, then glyph scaling; past all three its word spaces take the rest. Values are clamped to their range, and Minimum ≤ Desired ≤ Maximum must hold in every paragraph afterwards, or nothing changes and the command fails.
• autoLeading (0 to 500, default 120): the leading of characters whose leading is Auto, as a share of their size.
• singleWordJustification ("fullJustify" (default), "alignLeft", "alignCenter" or "alignRight"): how a justified line holding a single word is set.

The options are saved with the paragraph (justification in para) only when they differ from the defaults, and paragraph styles carry them (paraStyle.new {attrs: {justification: {auto_leading: 150}}}). Vertical type isn't glyph scaled, and type on a path doesn't use them. One undo step. ui.justification opens the Justification dialog on the values the Paragraph panel shows, with Preview; OK sends the values that changed.

{"name":"run_command","arguments":{"command":"text.setFormat","params":{"wordSpacing":[80,100,150],"letterSpacing":{"max":5},"singleWordJustification":"alignLeft"}}}

Kinsoku Set

text.setFormat {kinsoku: "hard"|"soft"|"none"} sets the Paragraph panel's Kinsoku Set of the selected type (or ids): which Japanese characters may not start or end a line. Hard (the default; new type and documents from before the setting) keeps closing brackets, commas, full stops, middle dots, iteration marks, the prolonged sound mark ー and small kana off the line start, and opening brackets off the line end. Soft lets 々, ー and small kana start a line, as JLREQ's level 3 line-breaking rules do (https://www.w3.org/TR/jlreq/#addendum_a3); the rest of the set stays. None applies no kinsoku. hard isn't saved. One undo step.

{"name":"run_command","arguments":{"command":"text.setFormat","params":{"kinsoku":"soft"}}}

Hanging punctuation (burasagari)

text.setFormat {burasagari: "none"|"standard"|"forced"} sets the Paragraph panel menu's Burasagari (None / Regular / Force) of the selected type (or ids): an East Asian comma or full stop ending an area type line (、。，．, half-width ､｡) hangs outside the frame, Regular only when it doesn't fit, Force always (the rest of the line then fills the measure). Closing brackets and Latin punctuation don't hang; point type, right-to-left and hyphenated lines are left as they are. The hanging mark is as wide as its punctuation spacing makes it (mojikumi). New type (the Type tools, text.create) takes standard; documents from before it and imported text keep none, and none isn't saved. One undo step.

{"name":"run_command","arguments":{"command":"text.setFormat","params":{"burasagari":"forced"}}}

Optical Margin Alignment

type.opticalMarginAlignment {on?: bool} (Type › Optical Margin Alignment) turns Optical Margin Alignment on or off for the selected type objects (or ids); without on it sets the opposite of the first object's setting. It replies {on, ids}, and the menu item is checked while the first type object it acts on (inside a selected group too) has it on. Optical Margin Alignment is a setting of the whole thread: the command sets every frame of a selected frame's thread, and ids lists them all. Threading frames gives the thread its first frame's setting, frames taken out of a thread keep it, and a file whose frames of one thread differ opens with the first frame's setting in all of them. In every left-to-right paragraph of point and area type, these characters hang outside the margin when they start or end a line: quotation marks, hyphens (a hyphenated line's hyphen too), full stops and commas hang wholly; asterisks, tildes, ellipses, en and em dashes, colons and semicolons hang by half their width; and A, V, W, Y, v, w and y hang by a tenth of theirs. A punctuation mark followed by a quotation mark hangs with it. Left-aligned lines hang at their left margin, right-aligned lines hang at their right margin, and centered and justified lines hang at both. In vertical type the margins are the top and bottom of a column. The space that justification adds between glyphs leaves out the gaps next to a hanging punctuation mark. Lines break where they did without it, and a line of one glyph does not hang. East Asian commas and full stops hang only through burasagari, and a line that ends in one hangs only at its start. Type on a path and right-to-left paragraphs are laid out as before. New type and documents from before it have it off, and it is saved only when on. One undo step.

{"name":"run_command","arguments":{"command":"type.opticalMarginAlignment","params":{"on":true}}}

Roman Hanging Punctuation

text.setFormat {romanHangingPunctuation: bool} (Paragraph panel menu › Roman Hanging Punctuation) turns it on or off for the paragraphs of the selected type (or ids, or the paragraphs a start / end range touches). In those paragraphs the punctuation Optical Margin Alignment hangs hangs outside the margin by the same amounts (quotation marks, hyphens, full stops and commas wholly; asterisks, tildes, ellipses, en and em dashes, colons and semicolons by half), on the same sides (left-aligned lines at their left, right-aligned at their right, centered and justified at both), and letters don't hang. With Optical Margin Alignment on as well, the letters hang too. East Asian punctuation hangs only through burasagari; type on a path and right-to-left paragraphs are laid out as before. Off by default, saved with the paragraph (roman_hanging in para) only when on, carried by paragraph styles, and turned off by Reset Panel. One undo step. The menu item is checked while the paragraph the Paragraph panel shows has it on.

{"name":"run_command","arguments":{"command":"text.setFormat","params":{"romanHangingPunctuation":true}}}

Small Caps, Superscript and Subscript

text.setFormat {smallCaps, position} (the selected type, or ids) and text.setRangeStyle {id, start, end, smallCaps, position} (a range) set what the Character panel's Small Caps, Superscript and Subscript buttons and menu items set. smallCaps: true draws lowercase letters as capitals at Document Setup's small caps size (70% of the font size by default); capitals keep their size, and with All Caps on every letter is a full-size capital. position: "superscript" or "subscript" draws the characters at Document Setup's superscript or subscript size, raised or lowered by its position (by default 58.3% of the font size, moved by 33.3% of it), and "normal" puts them back on the baseline. A file whose characters, character styles or graph label styles hold a superscript or subscript size outside 1 to 500%, a position outside -500 to 500% or a small caps size outside 1 to 200% opens with the nearer end of each range. Setting superscript on subscript characters replaces the subscript, and setting subscript on superscript characters replaces the superscript. When document.setup {superscript, subscript, smallCapsSize} changes the proportions, the characters take the new ones. One undo step. {"name":"run_command","arguments":{"command":"text.setFormat","params":{"position":"superscript","smallCaps":false}}}

Character Alignment

text.setFormat {charAlign} (the selected type, or ids) and text.setRangeStyle {id, start, end, charAlign} (a range) set where characters smaller than the largest on their line line up with it: romanBaseline (the default), the em box's emBoxTop, emBoxCenter or emBoxBottom, or the ideographic character face's (ICF) icfTop or icfBottom. In vertical type top and bottom are the right and left of the column. The ICF comes from the font's BASE table (icfb, icft, as the OpenType baseline tags define them), else from the average ink box of some ideographs and kana; a font without ideographs uses its em box. One undo step.

{"name":"run_command","arguments":{"command":"text.setFormat","params":{"charAlign":"icfTop"}}}

Proportional Metrics

text.setFormat {proportionalMetrics: true} (the selected type, or ids) and text.setRangeStyle {id, start, end, proportionalMetrics} (a range) set full-width glyphs on the proportional widths the font gives them with the OpenType palt feature; glyphs the font doesn't re-space keep their full width. In vertical type, upright glyphs take the proportional heights of the font's vpal feature (the glyphs are shaped top to bottom for it) and move with them; a font without vertical metrics keeps its cells. Line-end Punctuation Half Width then takes nothing more off the punctuation palt or vpal re-spaced (the opening bracket at a line's start, consecutive punctuation, the closing mark at a line's end); punctuation left full width is trimmed as before. With automatic kerning the font's kerning for the re-spaced glyphs comes too: apkn with palt, and vkrn and vapk with vpal (vertical type is otherwise not kerned). Off by default and saved only when on. One undo step.

{"name":"run_command","arguments":{"command":"text.setFormat","params":{"proportionalMetrics":true}}}

Variable fonts

text.setFormat {fontAxes: {"wght": 650}} (the selected type, or ids) and text.setRangeStyle {id, start, end, fontAxes} (a range) set variation axes of a variable font, as the Character panel's Variable Font sliders do: each tag named takes the value given, in the axis's own units and clamped to its range, in place of the style's setting of that axis, and null sets it back to the style's setting. The other axes keep their settings. A tag the font has no axis for is left out, and a value equal to the style's setting is not kept; for a font that isn't installed the values are kept as given. Layout, outlines, the substituted-glyph highlight and PDF export use the font instanced at the settings; SVG export writes them as font-variation-settings (and the wght setting as the weight, the wdth setting as font-stretch) and embeds the font instanced there, with a warning when embedded faces of one weight, style and width share characters; SVG import reads font-variation-settings back. Another font or style (text.setStyle {font, style}, text.setRangeStyle {font, style}, text.replaceFont, and a character style applied or redefined with another font unless it sets them) clears the settings, and Missing Glyph Protection keeps them on the characters that keep their font. They are saved as fontAxes only when set, and a character style captured from type (charStyle.new, charStyle.redefine) sets them explicitly, {} for none. Each command is one undo step.

{"name":"run_command","arguments":{"command":"text.fontList","params":{"family":"Bahnschrift","style":"Regular"}}}
{"name":"run_command","arguments":{"command":"text.setFormat","params":{"fontAxes":{"wght":650,"wdth":87.5}}}}

Digits

text.setFormat {digits: "default"|"arabic"|"hindi"|"farsi"} (the selected type, or ids) and text.setRangeStyle {id, start, end, digits} (a range) set the numeral system the characters' digits are drawn in, as the Character panel menu's Digits does with the Middle Eastern (Indic) options shown: arabic draws every digit as the Western 0–9, hindi as the Arabic-Indic ٠–٩ (U+0660–U+0669) and farsi as the Eastern Arabic-Indic ۰–۹ (U+06F0–U+06F9), whichever of the three sets was typed; default leaves each digit as typed. The text keeps the digits that were typed (text.getRange, search, the caret and selection are unchanged); only the glyphs change, as with All Caps, and SVG export writes the digits as drawn. A digit the font lacks comes from a fallback font. Default and saved only when set. One undo step.

{"name":"run_command","arguments":{"command":"text.setFormat","params":{"digits":"hindi"}}}

Touch Type

The Touch Type tool (touchType, Shift+T) changes one character at a time. A click on a character of a type object (pointer_gesture with touchType) selects the object and that character and shows a box around the character. Dragging the box or its lower-left handle moves the character along its line and up or down. The upper-left handle scales it vertically, the lower-right handle scales it horizontally, the upper-right handle scales it proportionally, and the handle in the middle of the top edge rotates it about its pen position on the baseline. Escape lets go of the character. Each drag is one text.touchType {id, start, dx?, dy?, hScale?, vScale?, rotation?} and one undo step. start is the byte offset of the character; a ligature (one glyph for several characters) changes as a whole, and the reply's start and end give its bytes. dx (document points, in reading order) changes the kerning of the character before it on its line and its own kerning, so the character moves by dx and the characters after it stay where they are. The first character of a line and a right-to-left character take no dx, and dx is limited to what kerning from -1000 to 10000 (1/1000 em) on the two characters can give. dy raises it by baseline shift, which moves a rotated character along its own up axis. The tool splits a drag of a character rotated by up to about 84 degrees between kerning and baseline shift, so the character follows the pointer; nearer a quarter turn, a drag moves it only along its line. hScale, vScale and rotation set its scales and rotation as text.setRangeStyle does. Where the font kerned the character with its neighbors, the kerning of the two characters keeps it and the characters after it where they were along the line; a wider or narrower character moves the characters after it by the change in its width. Touch Type gives the character before it a kerning value of its own (with dx, and with any other change where the font kerned the character with its neighbors), and that character can then move by the font's kerning pair with the one before it, as with any style change to one character (text.setRangeStyle). The type stays editable.

{"name":"run_command","arguments":{"command":"text.touchType","params":{"id":7,"start":1,"dx":10,"dy":4}}}

New type in a Japanese interface

While the interface is in Japanese (Preferences › User Interface › Language, or auto on a Japanese system), new type starts with em box top-to-top leading (leadingModel: "emBoxTop") and em box centre character alignment (charAlign: "emBoxCenter"); in other languages it starts on the Roman baseline (#432). This covers the Type tools, text.create and text.createInPath, which also take leadingModel and charAlign to choose, and paraStyle.new / charStyle.new made with no attrs and no text selected. The values used are kept in the journal, so a replay sets the same in any language. Saved documents keep what they say, and imported text (PDF, SVG, DXF…) keeps the Roman baseline.

{"name":"run_command","arguments":{"command":"text.create","params":{"x":40,"y":60,"text":"雅楽","leadingModel":"emBoxTop","charAlign":"emBoxCenter"}}}

Moving and flipping type on a path

Type on a path flows between a start and an end bracket, stored as fractions of its path's length. type.pathOptions {start?, end?, flip?, effect?, alignToPath?, spacing?, ids?} sets them and the rest of Type on a Path Options: end: null puts the end bracket back at the end of the path (or once round a closed one); flip then turns the type to the other side of its path, the path running the other way and the brackets swapping ends so the type keeps its stretch of the path; alignToPath runs the ascender, descender, center or baseline (the default) along the path; spacing (points) closes glyphs up round the outside of curves and opens them up round the inside. With none of them it queries the first selected type on a path: {start, end, flip: false, effect, alignToPath, spacing}. The Options dialog shows Effect, Flip, Align to Path and Spacing and leaves the brackets where they are. As in the reference app, the Selection and Direct Selection tools show selected type on a path's brackets: dragging the start or end bracket (pointer_gesture) sets where the type begins or ends, dragging the centre bracket slides the type along its path and, dragged across the path, flips it (with Cmd/Ctrl held it only slides). Each is one undo step.

{"name":"run_command","arguments":{"command":"type.pathOptions","params":{"start":0.25,"end":0.75,"flip":true}}}

Paragraph attributes per paragraph

Each paragraph (text split at \n) has its own alignment, indents, space before and after, hyphenation, punctuation spacing (mojikumi), hanging punctuation (burasagari), direction, leading model, tab stops and paragraph style. text.setStyle {justify}, text.setFormat {leftIndent, rightIndent, firstLineIndent, spaceBefore, spaceAfter, hyphenate, mojikumi, burasagari, direction, leadingModel}, text.tabs.set / text.tabs.clear and paraStyle.apply take optional start / end byte offsets of the plain text (as text.setRangeStyle does): they change the paragraphs the range touches (a caret, start == end, touches its paragraph). Without a range they change every paragraph of the targeted objects, as before. With a range, the character attributes of text.setStyle and text.setFormat style that range of characters. text.tabs.get {start?} and paraStyle.new / paraStyle.redefine {id, start?} read the paragraph at start; paraStyle.list counts paragraphs as uses. A Return the Type tool types (or text.editRange inserting \n) continues the style of the paragraph it splits; deleting a break keeps the first paragraph's style; text.setText gives every paragraph the first one's attributes. Threaded text keeps each paragraph's attributes as the story re-flows (a paragraph split between frames has them in both).

In the native file a text object's para holds the first paragraph's attributes and paras (left out when every paragraph is alike) one entry per paragraph, so older readers keep the first paragraph's. SVG export anchors each centred or right-aligned line on its own (<tspan text-anchor>), and SVG import gives each line's paragraph its alignment back, with indents that keep it in place. When a native file's type or character styles hold a size, a horizontal scale or a vertical scale of 0 or less, the value opens as 12 pt or 100%. A negative leading opens as Auto, and a size, leading or scale past 4,000,000 opens as 4,000,000. A baseline shift, an indent or a space before or after past 4,000,000 pt either way, in the type or in character or paragraph styles, opens at that limit, and a tab stop outside 0 to 100,000 pt opens at the nearer end.

{"name":"run_command","arguments":{"command":"text.setStyle","params":{"id":42,"justify":"center","start":0,"end":0}}}
{"name":"run_command","arguments":{"command":"text.setFormat","params":{"ids":[42],"spaceBefore":6,"start":12,"end":30}}}
{"name":"run_command","arguments":{"command":"paraStyle.apply","params":{"name":"Heading","id":42,"start":0,"end":0}}}

Constrain proportions

The link between W and H in the Transform panel, the Properties panel and the Control bar is the constrainProportions preference (prefs.get / prefs.set). With it on, those fields send proportional: true to object.setBounds, which scales the other dimension by the same factor; agents pass proportional themselves.

{"name":"run_command","arguments":{"command":"prefs.set","params":{"key":"constrainProportions","value":true}}}
{"name":"run_command","arguments":{"command":"object.setBounds","params":{"width":200,"proportional":true}}}

Plug-ins

WebAssembly plug-ins (plugins.md) are reached through commands; the desktop app and an unconfined vectorcraft-cli keep the ones plugin.install installs in a Plug-ins folder next to the preferences and install them again at startup (plugin.remove deletes them there): plugin.install {path | dataBase64} installs one, plugin.list / plugin.info {id} describe them (parameters, schema, defaults). An object filter runs on the selected paths and compound paths with plugin.run {id, params} (one undo step; it may change, delete or add objects, and the result is selected). A live-effect plug-in is an effect with id plugin.<id>: effect.apply, effect.setParams and effect.expandAppearance work as for built-in effects, and effect.list lists it. An exporter adds a write-only file format plugin.<id> (listed by document.formats): document.export {format: "plugin.<id>", params?, path?} writes it, as does a path ending in one of its extensions that no built-in format has. A failing plug-in returns an error and leaves the document as it was (an export writes no file).

{"name":"run_command","arguments":{"command":"plugin.install","params":{"path":"/plugins/desaturate.wasm"}}}
{"name":"run_command","arguments":{"command":"plugin.run","params":{"id":"org.vectorcraft.example.desaturate","params":{"amount":60}}}}
{"name":"run_command","arguments":{"command":"effect.apply","params":{"effect":"plugin.org.example.wobble","params":{"size":4}}}}

Colour adjustments

Effect → Color Adjustments holds live effects that recolour what an object paints with (fills and strokes, gradient stops, type, gradient meshes, embedded images; on a group or layer every member), each colour keeping its colour model: adjust.brightnessContrast, adjust.curves (points as "x,y x,y …" or [[x, y], …], 0–255), adjust.levels, adjust.hueSaturation, adjust.shiftToColor and adjust.temperatureTint (effect.list documents their parameters). Apply them with effect.apply like any effect, or with item to recolour one fill or stroke; they stay editable (effect.setParams), export with the adjusted colours (an adjusted image as a recoloured copy) and become permanent with effect.expandAppearance. Patterns and the shadows and glows of raster effects keep their colours.

Vector halftone

object.vectorHalftone turns the selection into vector dots, lines or shapes sized by its tone (mono, or CMYK screens multiplied over each other), clipped to the art's outline by default. An image in the art that can't be decoded is sampled as blank (no dots, or full dots with invert), and the result carries the warning that Flatten Transparency gives:

{"name":"run_command","arguments":{"command":"object.vectorHalftone","params":{"shape":"circle","frequency":25,"angle":45,"mode":"cmyk"}}}

Fonts

Type can use the bundled fonts, fonts added to the session and the fonts installed on the system (none on the web): the system's and the user's font folders (Windows: Fonts and %LOCALAPPDATA%\Microsoft\Windows\Fonts, plus fonts registered outside them, such as fonts installed as shortcuts, and in the desktop app and vectorcraft-cli (MCP included) the fonts in DirectWrite's system font collection, such as those Adobe Fonts activates while Creative Cloud runs; macOS: /System/Library/Fonts, /Library/Fonts, /Network/Library/Fonts, ~/Library/Fonts and downloaded system fonts, plus in the desktop app and vectorcraft-cli the fonts CoreText's font manager lists outside them, such as those apps and font managers register from their own folders; Linux and BSD: /usr/share/fonts, /usr/local/share/fonts, ~/.fonts and the XDG data folders' fonts, ~/.local/share/fonts among them, in a Flatpak sandbox the host's fonts, and the folders fontconfig's configuration lists (<dir> in /etc/fonts/fonts.conf and the files it includes: where packages add theirs, such as TeX fonts under /usr/share/texmf/fonts)). Faces without outlines VectorCraft draws (no glyf, CFF, CFF2 or VARC table, such as bitmap-only fonts) are left out. The scan also reads the folder the fontsFolder preference names (Preferences › Type › Additional Fonts Folder) with its subfolders and, in the desktop app and vectorcraft-cli, VectorCraft's own Fonts folder next to the preferences: ~/Library/Application Support/VectorCraft/Fonts on macOS, %APPDATA%\VectorCraft\Fonts on Windows, and $XDG_CONFIG_HOME/vectorcraft/Fonts or ~/.config/vectorcraft/Fonts on Linux and BSD. text.addFontFiles copies fonts into that folder. The installed fonts are cataloged once per session (in the background when the app starts, else on the first lookup by family name), so opening, placing, pasting and importing files find them whatever ran before. text.fontList lists every family available, the installed ones included, each by its English (canonical) name — the same string text.setStyle {font} and documents use — without the system's hidden families, whose names start with "." (macOS's ".SF NS", ".LastResort"), which still resolve when a document names them. Preferences › Type › Show Font Names in English (fontNamesInEnglish) only changes the UI menu labels, not this list. With family it gives that family's styles (upright by weight, then italics) and fails when the family isn't available; axes lists a variable font's variation axes (tag, name, min, default, max, hidden, and value, the setting of style, else of the first style), empty for a static font. text.rescanFonts (Character panel menu › Refresh Font List) scans the font folders again, for fonts installed or removed since the app started: type set in a font that became available redraws in it, and type set in a family that was removed redraws in the fallback font and is listed as missing, without editing the document. The app does so by itself when its window comes to the front and a font folder changed meanwhile.

{"name":"run_command","arguments":{"command":"text.fontList","params":{}}}
{"name":"run_command","arguments":{"command":"text.fontList","params":{"family":"Source Serif 4"}}}
{"name":"run_command","arguments":{"command":"text.rescanFonts","params":{}}}

text.missingFonts lists the fonts that the active document's type uses, in its layers and its symbols, and that aren't available as named: {fonts: [{family, style, status, resolved}], count, fontsNextToDocument?}. status is missing or substitute, as in text.fonts. CSS generic families such as sans-serif, which no font file provides, and names longer than 256 bytes are left out. fontsNextToDocument is the Fonts folder next to the saved document (File › Package writes it), when there is one and a search may start there. text.findFontFiles {folder} looks for the fonts' files in an absolute folder and its subfolders on separate threads and returns at once with state: "searching" and an id; fonts: [{family, style?}] names other fonts to look for. text.findFontFiles {} returns the last search's state with the files found so far for each font: searching, done, stopped (with stopped: stop, time or limit) or failed (with error); before the first search it returns {state: "idle", fonts: []}, without an id. {stop: true} stops the search. A search reads the table directories and the name, fvar and OS/2 tables of .ttf, .otf, .ttc and .otc files and, on macOS, of suitcase fonts (extensionless or .suit files whose fonts are in their resource fork), and it lists a file for a font only when adding the file makes the font resolve exactly as text.fonts resolves it. It ends after maxSeconds (60 by default, 1 to 600), at its limits on the entries listed (10,000,000), the folders waiting to be listed (1,000,000), the font files read (10,000) and the files kept (1,000), or once every font has a file. Folders more than 64 levels below the picked one are skipped and counted in skipped. A search does not read a font file larger than 256 MB, which text.addFontFiles does not copy, or a suitcase font whose resource fork is larger than 8 MB.

A search does not enter app and media library packages (.app, .bundle, .framework, .photoslibrary and the like), folders named Program Files (also Program Files (x86) and Program Files (Arm)), ProgramData, $Recycle.Bin or System Volume Information, the system's folders at the root of a volume (a macOS, Linux, BSD or Windows root, told by the folders it holds), the Shared and Public folders in Users, a home folder's Library, AppData, Applications and snap folders and the folders in it whose names start with a dot, the app data folders in a Library folder (Application Support, Containers, Group Containers, Caches, Preferences), or the folders the environment names for apps and the system (such as APPDATA, LOCALAPPDATA, ProgramFiles, SystemRoot and XDG_CONFIG_HOME), and a search of a folder inside one of them fails. The walk also skips other hidden folders and folders whose contents are in the cloud, and it follows no links; a picked folder may be one of those, and a picked path that goes through a link is searched at its target. Programs installed in other folders, such as a folder on another drive, are searched as any other folder is.

text.addFontFiles {files} copies files that the session's last search found into VectorCraft's Fonts folder, then scans the fonts again, and type set in them redraws. It copies only regular .ttf, .otf, .ttc, .otc, .woff and .woff2 files and suitcase fonts (with their resource fork), at most 256 MB a file and 1 GB in all, and lists any other file in skipped. It never replaces a file: the names Name.otf, Name 2.otf up to Name 99.otf are tried in turn until one is free, which gets the copy, or is a file with the same contents, which is kept (kept). Copy only fonts you own or are licensed to install. command.batch returns before a search ends; later requests over MCP or the control channel poll text.findFontFiles {}. vectorcraft-cli run stops the search when its last step ends. The web build has no Fonts folder and no folder search.

{"name":"run_command","arguments":{"command":"text.missingFonts","params":{}}}
{"name":"run_command","arguments":{"command":"text.findFontFiles","params":{"folder":"/Users/me/Downloads"}}}
{"name":"run_command","arguments":{"command":"text.findFontFiles","params":{}}}
{"name":"run_command","arguments":{"command":"text.addFontFiles","params":{"files":["/Users/me/Downloads/Example/Example-Regular.otf"]}}}

Tool options

tool.setOption reads and sets a tool's options, headless too: {key, value} or {values: {…}} sets them on the active tool, tool names another one, and {} just reads them; the result is the tool's options. The options a tool keeps last across tool switches (and documents) and are saved with the preferences (toolSettings in prefs.get's full object), as the reference app keeps them: the Liquify tools' brush and tool options, Mirror & Cut's axis and side, Puppet Warp's mesh, the Symbolism brush, the line, grid, pencil, brush and eraser tools' options, polygon sides and star points, and the Flare Tool Options. The Liquify tools share one set of Global Brush Dimensions (width, height, angle, intensity), and the Symbolism tools one brush. Interaction state (pins, a reference point) starts afresh.

{"name":"run_command","arguments":{"command":"tool.setOption","params":{"tool":"twirl","values":{"width":60,"rate":90}}}}

tool.options {tool} opens what double-clicking the tool's button opens (its params doc lists every tool), in the desktop app. The Flare tool's are its Flare Tool Options (dialog flareOptions): diameter and pathLength (pt), opacity, brightness, growth, fuzziness, longest, rayFuzziness and largest (%), rays, rings, direction (°), and raysOn and ringsOn (the Rays and Rings checkboxes; off, the flare has none). OK keeps them (tool.setOption {tool: "flare", values}) and the next flare dragged out uses them; a click with the tool opens the same dialog, and its OK also draws a flare there. The Artboard tool's are the active artboard's Artboard Options (dialog artboardOptions; the Artboard tool's active artboard while it is in use, else the window's; its x and y are the artboard's top-left corner in document points, while the dialog shows X and Y at the reference point the Transform panel uses, counted from the ruler origin), a graph tool's the selected graph's Graph Type (an error without one), a Symbolism tool's the Symbolism Tools Options (dialog symbolismOptions: diameter in points, intensity and density 1–10, one brush for the eight tools) and the Magic Wand tool's its panel. ui.newArtboard (Alt-click on the Artboards panel's New Artboard) opens Artboard Options for a new artboard, with new: true and the name, position and size that artboard.new {} would give it. OK runs artboard.new with the fields and makes the new artboard the active one, and Cancel adds no artboard. artboard.new and artboard.setProps trim the white space around name, and a blank name gives the artboard its default name, "Artboard N" for the Nth artboard.

{"name":"run_command","arguments":{"command":"tool.setOption","params":{"tool":"flare","values":{"rays":24,"ringsOn":false}}}}

Perspective grid

The grid is document data (perspective.grid.set edits the model directly). perspective.grid.get reads it: grid (the model), define (the Define Grid fields), station (the viewer: x the centre of vision on the horizon y, distance from the picture plane, in points) and defined (false while the document uses the default grid). perspective.grid.define sets the grid from the Define Grid fields: kind, units, scale ([artboard, real world]), gridline, angle (the viewing angle, 0–90°), distance (the viewing distance), horizonHeight, thirdVp ([x right, y up] from the centre of vision), leftColor, rightColor, groundColor ("#rrggbb") and opacity (0–100). Lengths are real-world lengths in units drawn at scale; missing fields keep the grid's. Only what differs changes (an unchanged OK is no undo step): a new type, angle or distance moves the vanishing points around the station point, which stays put. In two- and three-point grids the viewer looks at the left plane at the viewing angle, so the vanishing points sit at x − distance/tan(angle) and x + distance·tan(angle).

{"name":"run_command","arguments":{"command":"perspective.grid.get","params":{}}}
{"name":"run_command","arguments":{"command":"perspective.grid.define","params":{"units":"inches","gridline":0.5,"angle":30,"distance":6,"horizonHeight":3}}}

Perspective grid presets: perspective.presets.list lists the built-in views ([1P-Normal View], [1P-Low View], [1P-High View], [2P-Normal View], [2P-Low View], [2P-High View], [3P-Normal View], [3P-Low View]; builtIn: true, fitted to the first artboard, protected), then the saved ones, each with its Define Grid fields. perspective.grid.preset {name} (or {kind}, that type's normal view) resets the grid to one, fitted to the first artboard. perspective.presets.save {name?, newName?, preset?, …fields} saves a preset: just {name} saves the document's grid (View → Perspective Grid → Save Grid as Preset). perspective.presets.delete {name}, perspective.presets.export {names?, path?} (a .vcperspective JSON file; without path it returns data) and perspective.presets.import {path? | data? | dataBase64?, replace?} manage them. Saved presets are kept with the preferences, and a grid whose fields are a preset's is named after it (name).

{"name":"run_command","arguments":{"command":"perspective.grid.preset","params":{"name":"[2P-Low View]"}}}
{"name":"run_command","arguments":{"command":"perspective.presets.save","params":{"name":"Street","units":"feet","scale":[1,48],"gridline":1}}}

Perspective grid view options (View → Perspective Grid; view state saved with the document, no undo step, each {on?: bool}, no param toggles, → {on}): perspective.grid.lock (Lock Grid: the grid's widgets can't be dragged), perspective.grid.lockStation (Lock Station Point: dragging one vanishing point turns the view around the station point, so the other one moves), perspective.grid.snap (Snap to Grid, on by default: perspective.draw lands the drawn corners, and perspective.move the nearer edge, on gridlines within a quarter cell; both take snap: false) and perspective.grid.rulers (Show Rulers: a ruler up the line where the planes meet, in the grid's units at its scale). Gridlines draw in the Define Grid colours at its opacity.

Perspective grid widgets: with the Perspective Grid tool (pointer_gesture), the vanishing points (Shift keeps the horizon; with Lock Station Point the other one swings), the horizon, the third vanishing point, the origin, the left and right ground-level points (24 px from the origin along each wall's ground line: they move the whole grid; Shift keeps one axis), the left and right extents (separate: extent and extentRight; Alt drags both, Shift goes by whole cells), the vertical extent and the cell size widget (on the line where the planes meet, a cell up, or more cells while they're small) reshape the grid; Lock Grid stops them. The Plane Switching Widget stays put in a corner of the document window (headless: the first artboard's top-left corner); a press on it picks the plane with any tool while the grid shows (choosing a perspective tool shows it, and perspective.grid.show hides it again with the tool still chosen), and the keys 1–4 (key with "1"…"4") pick the left, horizontal, right and no plane. perspective.widget.options {show?, position?} (Perspective Grid Options, double-click the tool) hides it or moves it to topLeft, topRight, bottomLeft or bottomRight, kept with the preferences (prefs.get/prefs.set key perspectiveWidget).

Envelopes

Object → Envelope Distort wraps the selection in a live envelope: object.envelope.makeWithWarp (style: arc, arcLower, arcUpper, arch, bulge, shellLower, shellUpper, flag, wave, fish, rise, fisheye, inflate, squeeze or twist; bend, h and v in % from -100 to 100; horizontal or orientation), object.envelope.makeWithMesh (rows, cols: 1–50) and object.envelope.makeWithTopObject (the topmost selected path becomes the envelope). object.envelope.info reads the selected envelope's settings (type, the warp's or mesh's values, fidelity and the options, editing), or with no envelope selected the options new envelopes get. With an envelope selected, object.envelope.resetWithWarp and object.envelope.resetWithMesh (rows, cols, maintainShape: true keeps the current shape) switch its kind and keep its content. object.envelope.options sets fidelity and the options (antiAlias, preserveShape: clippingMask or transparency, distortAppearance, distortLinearGradients, distortPatternFills; the last two need distortAppearance); with no envelope selected it sets them for new envelopes. New envelopes start with Distort Appearance on, as in the reference app; envelopes saved before these options existed keep their old look (appearance applied after the distortion).

object.envelope.release gives back the content and the envelope's shape (a grey gradient mesh for warp and mesh envelopes, the path for a top-object envelope); the content keeps the envelope's opacity, blend mode and opacity mask (a group around it carries them when there are several objects). object.envelope.expand, and object.expand with object on, replace envelopes by groups of their distorted content that keep the name, transparency, knockout and opacity mask. Type inside an envelope distorts as its glyph outlines, run by run in each run's paint, on the canvas and in every export (SVG, PDF, EPS, EMF/WMF, DXF). Everything inside bends with the envelope: images as a raster mesh warp (cut into pieces by Fidelity, each a clipped image under its own affine map; SVG shares the pixels between the pieces), symbol instances as their symbol's art, and with Distort Appearance strokes (as their filled outlines) and geometry effects; with it, Distort Linear Gradients bends linear gradients and Distort Pattern Fills the pattern tiles. A transformed envelope keeps its warp in its own frame (kind.frame in document.inspect; rotating a warp envelope turns the warp with it), and move and scale keep it square to the page.

{"name":"run_command","arguments":{"command":"object.envelope.makeWithWarp","params":{"style":"arch","bend":40}}}
{"name":"run_command","arguments":{"command":"object.envelope.resetWithMesh","params":{"rows":3,"cols":3,"maintainShape":true}}}
{"name":"run_command","arguments":{"command":"object.envelope.options","params":{"fidelity":80,"distortAppearance":true,"distortLinearGradients":true}}}
{"name":"run_command","arguments":{"command":"object.envelope.info","params":{}}}

Liquify tools

The Liquify tools (warp, twirl, pucker, bloat, scallop, crystallize, wrinkle) drag a brush over paths: the selected ones, or with nothing selected every path the brush passes over. A stroke is one undo step and one object.liquify journal entry, which an agent can also run directly. Their options (tool.setOption) are the Global Brush Dimensions shared by the seven tools (width, height in pt, angle, intensity 0..1, usePressure, showBrush) and the tool's own: detail (1..10); simplify (0..100) with simplifyOn (Warp, Twirl, Pucker, Bloat); rate (Twirl, -180..180°); complexity (0..15) and affectAnchors, affectIn, affectOut (Scallop, Crystallize, Wrinkle: what the brush moves); horizontal, vertical (Wrinkle, 0..1). Alt-drag sizes the brush from its current size (with Shift too it keeps its proportions). With Use Pressure Pen on, each pointer_gesture event's pressure (0..1) is the intensity there, and object.liquify takes points as [x, y, pressure].

{"name":"run_command","arguments":{"command":"tool.setOption","params":{"tool":"bloat","values":{"width":80,"height":80,"usePressure":true}}}}
{"name":"pointer_gesture","arguments":{"tool":"bloat","events":[{"kind":"down","x":280,"y":150,"pressure":0.3},{"kind":"drag","x":280,"y":200,"pressure":0.8},{"kind":"up","x":280,"y":250}]}}
{"name":"run_command","arguments":{"command":"object.liquify","params":{"tool":"scallop","points":[[300,150],[300,250]],"complexity":3,"affectAnchors":false}}}

Holding the brush still keeps Twirl, Pucker and Bloat working, scaled by the time held: every tenth of a second the stroke's last point repeats in object.liquify's points, one more dab there (other tools ignore repeated points). Give a pointer_gesture event holdMs (0..60000) to hold the pointer that long after it, button down; the same time gives the same result. Liquify reshapes paths only: type, symbols, images, graphs, meshes, and envelopes, repeats and blends with their contents stay as they are, and guides are never touched. object.liquify returns those under the brush in skipped with a warning, which the gesture's requests carry as {"status": …} (the status bar in the app). A drag applies only the dabs each new sample adds; the result equals applying the whole stroke at once, so the journal entry replays it exactly.

{"name":"pointer_gesture","arguments":{"tool":"twirl","events":[{"kind":"down","x":300,"y":200,"holdMs":800},{"kind":"up","x":300,"y":200}]}}

Perspective Selection

Objects in perspective keep their attachment themselves (perspective on the object: plane and depth, the distance along the plane's normal), so copies, duplicates and pastes stay attached. With the perspectiveSelection tool, a drag moves the selection within its plane (perspective.move; Alt copies), the bounding-box handles scale it in plane space (perspective.transform; Shift keeps proportions, Alt scales about the centre), and press_key 5 during a drag switches the move to perpendicular to the plane (press again to switch back). The arrow keys nudge a selection in perspective by the keyboard increment (perspective.nudge; Shift ×10, Alt copies). Object › Transform › Transform Again repeats the last perspective move or scale in plane space.

{"name":"pointer_gesture","arguments":{"tool":"perspectiveSelection","events":[{"kind":"down","x":480,"y":410},{"kind":"drag","x":470,"y":405}]}}
{"name":"press_key","arguments":{"key":"5"}}
{"name":"pointer_gesture","arguments":{"events":[{"kind":"up","x":470,"y":405}]}}
{"name":"run_command","arguments":{"command":"perspective.transform","params":{"matrix":[1.5,0,0,1.5,0,0],"copy":true}}}
{"name":"run_command","arguments":{"command":"object.transformAgain","params":{}}}

Planes move along their normals: perspective.plane.move {plane, offset | by, objects} (leftOffset, rightOffset, groundOffset in the grid). The plane widgets (a diamond on each plane, drawn while the grid shows) do it with the Perspective Grid and Perspective Selection tools: a plain drag moves the plane alone, Shift-drag moves the objects on it too, Alt-drag copies them; double-clicking a widget (or ui.perspectivePlane {plane}) opens the perspectivePlane dialog (location in points, objects: none, move or copy). Objects keep their place when the grid's definition changes, as in the reference app; perspective.grid.set with reproject: true moves them with it instead. perspective.plane.matchObject (Object › Perspective › Move Plane to Match Object) moves the plane of the selected object onto it.

{"name":"run_command","arguments":{"command":"perspective.plane.move","params":{"plane":"right","offset":40,"objects":"move"}}}
{"name":"run_command","arguments":{"command":"perspective.plane.matchObject","params":{}}}
{"name":"run_command","arguments":{"command":"ui.perspectivePlane","params":{"plane":"ground"}}}

Type and symbol instances attached to a plane stay type and symbol instances: their perspective record keeps a projection (a 3 × 3 matrix) the canvas and every export (SVG, PDF, EPS, EMF/WMF, DXF, PNG) draw their outlines through, so they are really foreshortened and still editable. perspective.editText (Object › Perspective › Edit Text, or double-clicking the type with the Perspective Selection tool) shows the type flat where it is drawn, in isolation mode, for the Type tool and text.* commands; object.exitIsolation projects it again. Release with Perspective keeps their look. Shape, spiral, polar grid and flare drags draw on the active plane while the grid shows, and so do the click-to-size dialogs (perspective.draw with at: the click; sizes are plane units).

{"name":"run_command","arguments":{"command":"perspective.editText","params":{}}}
{"name":"run_command","arguments":{"command":"text.setText","params":{"text":"OPEN"}}}
{"name":"run_command","arguments":{"command":"object.exitIsolation","params":{}}}
{"name":"run_command","arguments":{"command":"perspective.draw","params":{"command":"shape.rectangle","params":{"x":470,"y":300,"width":50,"height":30},"at":[470,300]}}}

Blends

object.blend.make blends the selected objects (or ids) into a live blend. Spacing and orientation default to what object.blend.options set with no blend selected (object.blend.info reads them: target: "defaults"), else Smooth Color and Align to Page. starts gives, per object, the anchor of its first subpath the blend starts from (as the Blend tool's clicks on anchor points do; an open path's last anchor runs it the other way). A blend among the objects takes the others in as more key objects, keeping its options, name and transparency:

{"name":"run_command","arguments":{"command":"object.blend.make","params":{"ids":[12,15],"starts":[0,2],"steps":8}}}
{"name":"run_command","arguments":{"command":"object.blend.make","params":{"ids":[20,31]}}}
{"name":"run_command","arguments":{"command":"object.blend.info","params":{}}}

A blend's spine is the straight lines between its key centres until it is edited. object.blend.info lists it (spine.anchors, keyAnchors: the point each key sits on). object.blend.spine.addAnchor {x, y} adds a point where the spine passes nearest, object.blend.spine.moveAnchor {anchor, x, y} moves a point (a key on it moves along; with handle: "in" | "out" it places that handle), object.blend.spine.convertAnchor {anchor, to: "smooth" | "corner", x?, y?} converts a point's handles (a Smooth drag with x, y creates opposite handles; Corner retracts both), object.blend.spine.reshapeSegment {segment, t, dx, dy, perpendicular?} bends a segment so its point at t moves by (dx, dy), turning a straight spine into a curve (perpendicular works as in path.reshapeSegment), and object.blend.spine.removeAnchor {anchor} deletes a point no key sits on; the first edit turns the spine into a path. Moving a key moves its end of the spine. Release (object.blend.release) leaves the spine as a path that paints nothing (spines in the result):

{"name":"run_command","arguments":{"command":"object.blend.spine.addAnchor","params":{"id":40,"x":150,"y":100}}}
{"name":"run_command","arguments":{"command":"object.blend.spine.moveAnchor","params":{"id":40,"anchor":1,"x":150,"y":20}}}

How blends interpolate: closed shapes start where they twist least unless starts picks the anchors; strokes interpolate weight, dashes (a solid stroke counts as the dashed one with its gaps closed), miter limit, arrowhead scale and width profiles, while caps, joins, alignment, arrowheads and brushes switch halfway; gradients of one kind with different stop counts resample to a common set of stops. Groups pair their members in stacking order (members only one group has grow out of the other's centre), compound paths stay compound paths, and type, symbol instances and images interpolate their transforms (type also its colours, stroke weight and size run by run; the words and the symbol switch halfway). New blends are knockout groups (object.setProps {knockout} changes it), drawn on the canvas as they export. object.blend.expand and object.blend.release keep the blend's name, transparency, opacity mask and appearance (Release on a group around the keys and spine when the blend has any), and object.expand with object: true expands the blends in the selection.

Objects on Path

object.objectsOnPath.attach {path?, ids?, pivot?, rotate?} attaches the selected objects (or ids) to the path path: the objects and the path go into an Objects on Path group, with the path at its bottom. The objects go along the path in their order on the canvas, column by column from left to right and top to bottom in each column. The first one sits on the path's first anchor, the last one sits on the end of an open path, and the others are spread evenly between them by length; on a closed path the gaps run all the way round. The attachment point of each object (pivot, 0 to 8 row by row from the top left of its bounds without strokes; default 4, the center) sits on the path, and the object is turned to the path's direction there. An object rotated by other than a multiple of 90° is first turned to the nearest multiple. rotate tilts every object by that many degrees counter-clockwise. Without path, the command chooses the Objects on Path tool (objectsOnPath), whose click on a path sends {path}. As the path, the command refuses a guide, a compound path, type, a path without length and a path already in an Objects on Path group. As objects, it refuses graphs, layers and objects already in an Objects on Path group.

The group stays live: after every edit its objects go back on its path. Editing the path moves them, an object moved on its own goes back to its place, an object moved into the group (layer.move with place: "inside") joins the end of the path, an object moved out of it leaves, and art that replaces an object (symbol.breakLink) keeps its place. object.objectsOnPath.options {ids?, pivot?, rotate?} changes Pivot and Rotate, and object.objectsOnPath.expand {ids?} makes the groups plain groups. Both act on the groups that ids names (Objects on Path groups or objects in them), or on the groups of the selection when ids is not given. The native file keeps the group with an objects_on_path record; SVG, PDF and the other exports write the path and the objects as plain art.

{"name":"run_command","arguments":{"command":"object.objectsOnPath.attach","params":{"path":30,"ids":[12,15,18]}}}
{"name":"run_command","arguments":{"command":"object.objectsOnPath.options","params":{"pivot":7,"rotate":15}}}
{"name":"run_command","arguments":{"command":"object.objectsOnPath.expand","params":{}}}

Paragraph composer

text.setFormat {ids?, composer: "singleLine"|"everyLine"} sets how a text object's lines are broken (the Paragraph panel menu's Single-line and Every-line Composer; stored in the paragraph attributes as para.composer, saved only when it is Single-line). Every-line, the default, picks the breaks of the whole paragraph together in area type: justified text gets even word spacing, ragged text (left, centre or right aligned) an even rag, so it may move a word that fits to the next line to avoid a lone short word on the last line or a full line next to a short one. Single-line fills each line as far as it goes. Point type and type on a path have no line width, so both composers give the same result there; paragraphs that mix type sizes fall back to Single-line.

{"name":"run_command","arguments":{"command":"text.setFormat","params":{"ids":[7],"composer":"singleLine"}}}

Editing envelopes

object.envelope.editContents {editing} switches between editing the envelope and its contents (the menu item reads Edit Envelope while the contents are edited; it has no default shortcut, since the reference app's is Paste in Place here). While the contents are edited, clicks hit the content where it sits undistorted and the Selection tool selects it. A mesh envelope's points have bezier handles (kind.handles, offsets right/left/down/up; none stored means the smooth mesh through the points, as older files have): object.envelope.setMeshPoint {id, index, x, y, handle?} and object.mesh.movePoint, object.mesh.addLine and object.mesh.deletePoint edit them like a gradient mesh's (one undo step each). On the canvas a selected envelope shows its mesh; the Mesh tool and Direct Selection drag its points and the clicked point's handles, and the Mesh tool adds a row and a column where it clicks inside. The Control bar shows Edit Envelope / Edit Contents, the warp's style, orientation, bend and distortions (or the mesh's rows and columns), Reset (an unbent warp, a flat mesh) and Envelope Options.

{"name":"run_command","arguments":{"command":"object.envelope.editContents","params":{"editing":true}}}
{"name":"run_command","arguments":{"command":"object.mesh.movePoint","params":{"id":12,"index":4,"x":220,"y":140,"handle":0}}}

Shortcut sets from other apps

shortcutSet.read {path? | data? | dataBase64?, name?} reads a shortcut set another app saved as text (a .kys file that starts with /Menus {) without changing the document or the preferences. Each block that sets /Key is an entry, listed with its name, its section (the names of the blocks around it), its codes as the file writes them (key, modifiers, context, represent) and its line. codes counts the entries that use each key and modifier code, and unread lists the tokens the reader skipped, with their lines. The reader stops at 4 MiB, 16 open blocks, 20,000 entries and 8 MiB of names over the entries and their sections, cuts names at 1,024 bytes, and sets truncated when it reaches one of these limits. An XML shortcut file fails with "This shortcut file is in a form VectorCraft can't read."

The import matches record what the names and codes in such files stand for in VectorCraft. They are kept with the preferences, and resetting the preferences keeps them. importMatches.set {matches: [{kind, name, target, label?}]} records them: kind name maps an entry name to a command id, tool:<id> or panel:<id>, and kinds key and modifiers map a code, written in decimal, to a key or to modifiers as Edit › Keyboard Shortcuts writes them (K, Cmd+Shift). A target of "" records that there is no equivalent (for kind modifiers, no modifiers), and null removes the match. importMatches.list {kind?, query?} lists them and importMatches.reset {kind?} removes them. shortcutSet.read then adds the saved target to each entry, a chord to each entry whose codes are both known, and the saved key or mods to each code.

In the desktop app, Edit › Keyboard Shortcuts › Import… and File › Open show such a file in Import Shortcut Set (ui.shortcutImportDialog {path? | dataBase64?, name?} over the control channel). The dialog proposes targets from the entry names and decodes the codes that VectorCraft's default shortcuts agree on.

{"name":"run_command","arguments":{"command":"shortcutSet.read","params":{"path":"/Users/me/Studio.kys"}}}
{"name":"run_command","arguments":{"command":"importMatches.set","params":{"matches":[{"kind":"name","name":"Group","target":"object.group"},{"kind":"key","name":"901","target":"G"}]}}}