An override can't take its layer off the main component's text path, so an instance text of more than one line, whose main is on a path, stays on that path in the file. The export warned the general instance mismatch for it. It now says exactly that case: "an instance's text of more than one line stays on its main component's text path in the file, as an override can't take it off". FigCraft's own import reads such a layer as plain text. The docs' Left out list names the case.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove and wrote
* M6: .fig export writes a page's prototype background (#2326)
The export writes a page's prototype background as prototypeBackgroundColor (field 141, a plain colour) when it differs from the default, as Figma does for a page whose background was set. The device-settings warning now fires only for the device, rotation, colour, custom size and responsive settings. Measured on a synthetic file saved from Figma's web app.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Opus 5.5 briefed; Sonnet 5.5 measured and wrote at medium
* M6: the .fig export writes a prototype background opaque and warns for its opacity, and its tests read the written record (#2326)
The colour is written with alpha 1, as the measured record has it and as the import reads it, and a background that isn't opaque warns that its opacity was left out. The tests now read the written record, and each device setting alone is checked.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Fable 5.2 reviewed; Sonnet 5.5 wrote at medium
* M6: the .fig export writes every page's prototype background, so Figma shows the same colour (#2326)
A page at FigCraft's default background used to write no field, so Figma showed its own default instead. Every page now writes its colour, opaque. The test for a default page reads the written record and the import.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Opus 5.5 briefed; Sonnet 5.5 wrote at medium
---------
Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>
A curve's control points are no longer rounded on their own before their offsets from the end points are taken, so a curve read from a file is written back with the offsets it had. The instance test now reads both cases of the instance warning back.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Fable 5.2 reviewed; Sonnet 5.5 wrote at medium
Text on a path was written as an ordinary text layer under the warning "text on a path was written as plain text", so the path, where the text starts and its side were gone when the file was opened again. It is now written as the format's own layer kind (TEXT_PATH): the first part of its path as a vector network of one open or closed chain with the segment the text starts on marked, the start as that segment's own parameter (textPathStart.tValue, found by inverse arc length) and the side as forward. Vertical alignment is written as top, which is how text on a path draws.
More than one line and paths the format can't hold stay plain text, the parts of a path after the first are left out, and text on a path in an instance but not in its main component (or the other way round) is written as the main has it, each with its own warning. The field numbers come from the schema of a synthetic file saved from Figma's web app.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Fable 5.2 briefed; Sonnet 5.5 wrote at medium
* M6: .fig export writes list spacing, hanging punctuation and lists, and vertical trim (#2305)
The export wrote none of a text's list spacing, hanging punctuation, hanging lists or vertical trim, although the import reads all four. It now writes listSpacing, hangingPunctuation, hangingList and leadingTrim, each only when it differs from its default (as Figma saves them), and an instance override back to the default clears the field. The warning narrows to OpenType features, which the import does not read from a .fig.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Fable 5.2 briefed; Sonnet 5.5 wrote at medium
* M6: test that an instance override back to the default clears it (#2305)
Adds a round-trip test for the unset entries: an instance whose text goes back to the defaults must not take the main's list spacing, hanging flags and vertical trim again. Also scopes the comment and the docs row to a design file: a text at the defaults carries none of the four there, and a text style holds none of them.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Fable 5.2 briefed; Sonnet 5.5 wrote at medium
---------
Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>
Gap 1's 2026-10-26 bullet still said the export drops region fills.
tail_tests asserted on "region paints", which the narrowed warning can
never contain; it now checks for any region warning.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove and fixed; Fable 5.2 reviewed
Each vector region with paints of its own (an empty list included) now names an entry of the network's style table, so the paints come back on import; equal lists share one entry and a region without paints writes style 0. Only the paint frame of a region, which the importer cannot read, still warns.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Fable 5.2 briefed; Sonnet 5.5 wrote at medium
A document can hold image adjustments outside -1..=1 or not a number (a saved file, or a fills property); the renderer and the importer clamp them, but the export wrote the raw values for other readers. The export now writes the sanitized values, and a paint whose adjustments are all neutral after sanitizing writes no `paintFilter`. The round-trip test pins both cases and that all seven fields are written on a partly neutral paint.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Fable 5.2 briefed; Sonnet 5.5 wrote at medium
An image paint's adjustments (exposure, contrast, saturation, temperature, tint, highlights, shadows) were dropped on export with a warning, although the importer already reads them. The export now writes them as the paint's `paintFilter`: seven floats, with saturation as vibrance and the tint's sign flipped to the file's (positive towards magenta). A paint without adjustments gets no `paintFilter`. The warning is gone.
The schema gains `PaintFilterMessage` and `Paint.paintFilter`. A round-trip test turns the importer's fixture around and pins the written values and field numbers. Not checked in other software.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Fable 5.2 briefed; Sonnet 5.5 wrote at medium
The engine test now follows the issue: a list typed at 5, saved as .fig and opened, split by Backspace at the start of "b". The REST fixture is a top-level list, the junk sweep reaches the new loop through a listed text layer, and the document crate tests the new rule for nested, bullet, plain and over-deep levels. ParaList gets its doc comment back, and the revision row names the item written with 0.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Fable 5.2 queued; Opus 5.5 briefed; Sonnet 5.5 wrote at medium
The REST reader caps every size at 1,000,000; the native outline was built
from the uncapped size, so it reached past its frame.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove and fixed
Figma writes a list's start on every line of the list (measured in Figma's web app, 2026-11-12, from a saved .fig: a list started at 5 has listStartOffset 4 on all its lines, and the item split off by Backspace is written with 0). The import kept every line's value, so after a split the items after it restarted at the old start instead of 1.
The import now keeps a start only on the item where a numbered list starts counting, and gives every item that goes on counting 0, for .fig and REST listStartOffsets alike. The rule is one function in the document model, list_item_starts_counting, which the layout's list numbering now uses too. The export is unchanged and its re-import now equals what went in.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Fable 5.2 queued; Opus 5.5 briefed; Sonnet 5.5 wrote at medium
A native shape frame now writes fixed sizing on both axes, as a shape made here has, so the first edit no longer shrinks it to hug its text. A shape whose generated text is blank while the outer layer still carries text keeps the generated layers (#1397). The kind name in the no-outline note is clipped, the docs rows match the code, and the new key sits in order.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Fable 5.1 briefed; Sonnet 5.5 wrote at medium
On a whiteboard, a FigJam shape with text whose generated records are complete becomes a native board shape for the 23 kinds FigCraft draws: an empty centred frame holding FigCraft's own outline (the file's paints, weight and dashes kept) and the text. The 7 kinds without an outline, and shapes with damaged records, keep the generated layers; the former warn with the kind's name. Stickies and shapes share one native-board hook in the layer mapper.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Fable 5.1 briefed; Sonnet 5.5 wrote at medium
In Figma's data a component set holds one text property per variant (plus leftovers of deleted variants), and an instance lists only the ones its current variant's layers read (measured in Figma's web app). FigCraft listed every definition on every instance, and wrapped a long property name inside an 84 x 20 box so its lines painted over the rows around it.
- components::instance_prop_list: prop_list keeping slots and the properties a layer of the component references (a nested instance counts, its layers and slot content are not entered). component.properties and the Properties panel use it for instances; sets, components and variants still list every definition.
- prop_label cuts the name to one line with an ellipsis and shows the full name on hover.
- Tests: an_instance_lists_only_the_properties_its_variant_uses, a_long_property_name_stays_on_one_line, and the pin a_set_keeps_each_named_text_definition_and_layers_read_their_own. Two tests and one setup that listed an unbound boolean on an instance now bind it to a layer.
- docs/components-parity.md: the properties row, header and revision history.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Fable 5.1 briefed; Sonnet 5.5 wrote at medium
A boolean group takes a corner radius (and corner smoothing): the Design panel shows the field for it, `node.set cornerRadius` writes it, and the renderer rounds every sharp corner of the group's result after the fold, convex and concave alike. Layers inside keep their own radius and do not stack with it, a radius larger than the edges allow shrinks to fit while the field keeps the typed value, Flatten bakes the arcs, and a mixed selection reads Mixed. Measured in Figma's web app on 2026-11-10.
The rounded result is cached apart from the fold (keyed by the folded path, radius and smoothing), so scrubbing the radius never redoes the boolean steps. The .fig and REST importers and the .fig exporter now carry the radius and smoothing of a boolean group, and paste properties and radius variables reach it. Docs: ui-parity, target-app-parity, gaps, file-format-parity, ROADMAP, mcp.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Fable 5.1 designed; Sonnet 5.5 wrote at medium
From the pre-ready review: a sticky or shape item on a non-frame layer is ignored with the warning; the author drops control characters as board.sticky does; dump_samples writes a whiteboard's board items; a doc link made plain text.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove and wrote; Fable 5.1 reviewed
On a FigJam whiteboard the mapper builds a sticky with complete generated records as a native board sticky: the saved colour, size, place, opacity and shadows, the note and the author line as text layers, no background vector. It writes a board field on the REST JSON, which the REST importer reads (junk values are ignored with a warning). Damaged stickies and stickies in a .fig keep the generated layers.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Fable 5.1 briefed; Sonnet 5.5 wrote at medium
docs/ui-parity.md conflicted on the wrap style and text direction rows: kept this branch's text with #2135's Developer Mode naming.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove and wrote
Figma holds a US registration for DEV MODE (US 7574348). Display text only: the view.devMode label, Inspect tab, shortcut row, refusal messages, CLI help, docs and all locale catalogs now say Developer Mode. Command ids, preference keys and saved fields are unchanged; "Dev Mode" stays a search alias in the Actions palette. Adds a naming guard test.
Figma rounds every point whose style doesn't set a corner radius with the
layer's cornerRadius. The import read only the point's style, so those
vectors came in sharp. Points now fall back to the layer's radius, at the
scale of the layer that carries it (#2091), and the network cache keys on
it. The export writes an explicit 0 for a sharp point on a rounded vector
so it stays sharp.
bflatastic
Trevor BrownandClaude Opus 5.5
skylar taylor-barrick
Giovanni Almeida
FredApps