Introduce a leafer-ui canvas layer that mirrors rendered nodes and drives
viewport, node-drag, and alignment-guide updates imperatively, committing
React state on idle via startTransition to keep pan/zoom/drag at 60fps. Wire
the imperative viewport through wheel zoom, zoom-to, fit-view, and panning.
Add a dedicated crop tool: click a node to auto-detect a color-boundary crop
region or drag to preselect one, previewed with an on-canvas draft overlay
before entering the crop editor. Route uploads through renderContent so the
transparent WebP texture feeds the fast display path. Add node --test
coverage for the leafer renderer and color-boundary detection.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Rasterize imported SVG/CAD sources to a transparent WebP render texture and
carry it as an optional renderContent alongside the original content through
the domain node, snapshot normalization, and reference images. Add
canvasNodeRenderContent and prefer it for artboard, layer, generated-file,
and project thumbnail previews while keeping content as the crop/download/
model-reference source. Generalize the SVG rasterizer with a render purpose
and add upload/reference coverage.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add an optional renderContent field to canvas nodes across the go-zero API
type, domain node, SQLC models/queries, PostgreSQL schema, and mappers so a
lightweight transparent-WebP display texture persists alongside the original
content. Prefer renderContent for project thumbnails and main-canvas
previews, and treat both content and renderContent as protected/deletable
asset URLs during canvas asset cleanup. Extend round-trip coverage.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Extend CAD conversion to emit bounded semantic metadata (units, layers,
blocks, entity types, labels, crop-indexed bounds) alongside the derived
SVG, tag imported drawings as cad-vector nodes, and carry semanticContext
through canvas nodes, snapshot normalization, and reference images. Agent
Chat forwards this metadata as hidden cad-context reference parts.
Add non-destructive vector cropping: cropping a vector/CAD node keeps the
source SVG and spawns a transparent WebP sibling with crop-scoped semantic
context, wired through the vectorized image toolbar and the workspace crop
tool. Include node --test coverage for CAD semantics and vector crop.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add a semantic_context column to canvas_nodes (schema, sqlc queries,
generated models, repository upserts) and thread SemanticContext through the
domain node, API type, and mappers so CAD-derived context survives canvas
saves.
Treat agent text parts with textSource "cad-context" like settings: keep
them for planning, memory, and image-generation prompts but exclude them
from the persisted user-visible message copy. Add coverage for the CAD
context visibility and mapper behavior.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Move canvas wheel zoom off the React onWheel prop into a manually attached
non-passive listener so preventDefault reliably suppresses page scroll, and
read the viewport from the ref to keep the handler referentially stable.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Accept DWG and DXF files (up to 24 MB) through the canvas asset picker and
drag-and-drop. Conversion runs locally in a browser worker via acad-ts,
uploads a sanitized derived SVG, and leaves the source file untouched. DXF
covers ASCII and binary drawings; DWG covers AutoCAD R14 through 2018.
Rename the image-only upload input/handler to a generic canvas asset flow,
add vector-image node detection, tag imported SVGs as vector-image nodes,
localize CAD import status/error messages, and switch the Next build to the
webpack compiler so the worker bundles correctly. Add a node --test suite
covering CAD conversion and vector image helpers.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add modelImageReferences to preprocess agent contents, uploading local
image references and surfacing their names in the prompt so the model can
cite them. Wire the preparer into the canvas agent submission path, pass
imageName through the floating composer references, and relax the prompt
directive parser to accept names containing arbitrary characters.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Drive the drop-target border radius from --image-drop-radius, enlarge the
overlay radius, and widen the composer editor padding.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add a VisualAnnotationPanel letting users mark regions on an image (brush,
circle, rectangle, cross) with per-mark edit instructions, composited into
an annotation guide image. The new "visual-annotation" node action sends
the unmarked source plus the annotated guide to the image model, editing
only numbered regions. Server validates the annotation image/instructions
and threads them through the generator task.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Clarify in API/README that device limits are enforced independently per
normalized device type with atomic overflow revocation, and add a mapper
test covering the RFC3339 UTC timestamp contract.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add an ImageDropOverlay component and useImageDropTarget hook that let
users drop an image onto the home prompt card or canvas agent composer to
upload it as a reference, with a localized drop overlay.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Track authenticated sessions per device (parsing user-agent for desktop
vs mobile via mileusna/useragent), enforce configurable concurrent device
limits (Auth.DeviceLimits), and surface device status and "remove other
devices" management in the account dialog. Adds device/session context to
the auth module stores and exposes limits through the API.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Track agent history loading in the workspace and render a spinner with a
localized "loading conversation history" message in the assistant feed,
so the empty state no longer flashes before history resolves.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Condense the settings-chip summary to quality, ratio, and image count
(dropping resolution and pixel size), wrap the label in an ellipsized
span, and swap the up/down chevron for a single rotating ChevronDown
across the canvas agent and image generator composers.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Render a CanvasAgentEmptyState in the Agent panel that fetches localized
starter content from the new agent-starters endpoint, with local defaults
as fallback. Clicking a starter focuses the prompt composer (new focus
handle). Includes accompanying UI polish: side-nav cleanup, animated
prompt examples, and canvas panel spacing tweaks.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add a public GET /api/ui/agent-starters endpoint returning localized
empty-state content (title, description, preview images, and starter
prompts) for the canvas Agent panel. Content is server-configurable via
AgentStarter config with built-in zh-CN/en-US fallbacks, and the route is
exempt from authentication so shared canvases render the same content.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add a brand kit domain model, repository, and dedicated page/route for
managing brand kits, plus a BrandKitSelector used on the home page and
canvas workspace. Home project creation and the workspace can attach a
brand kit to a project, with new i18n strings and a side-nav entry.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add a brand kit domain with memory and postgres stores, a BrandKitService
for CRUD and resolution, and REST endpoints to list, upsert, and delete
brand kits. Projects can bind a brand kit whose resolved context and
reference images feed into the agent's long-term memory and design prompts.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Introduce sm/md/lg sizes on the Switch component driven by CSS custom
properties, move switch styling into a dedicated switch.css, and remove
the duplicated overrides from styles.css and ToolbarCustomizeDialog.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Move PSD generation off the server. The image-generation API only
produces raster formats, so drop the PSD request path, the oov/psd
decoder, and the LayeredDocumentGenerator wiring. Layer separation now
returns clean background, foreground, and text_render_data artifacts
into a transparent frame.
Export that frame (or any image node) as PSD from the canvas context
menu: ag-psd assembles layers, rotation/flip, opacity, editable text,
and stroke effects from the current canvas state at download time.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Replace the ratio-baked size presets with a resolution tier (1K/2K/4K) plus
ratio selection and normalized dimensions, send imageQuality/imageCount to
the agent, and add the output_format field to GeneratorTaskInputArgs.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Map imageQuality/imageCount and output_format across the HTTP and
WebSocket handlers, pass them into the design service calls, and register
the layered document generator in the service context.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add quality and output_format to ImageRequestOptions, plumb them through
the JSON/multipart request bodies, generalize asset URL collection to
non-image formats, and add GenerateLayeredDocument for PSD output.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Convert GPT Image 2 layered PSD output into canvas image/text nodes with a
local separation fallback (new layer_separation.go, oov/psd dependency),
default layer-separation/edit-elements output_format to psd, and thread
imageQuality/imageCount through project generation and job processing.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Add ImageQuality/ImageCount to AgentRequest, AgentChatRequest and Job,
add OutputFormat to GeneratorTaskInputArgs, and introduce the
LayeredDocument types and LayeredDocumentGenerator port for PSD output.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Extend CreateProject/Generate/AgentChat request types with imageQuality
and imageCount, and add output_format to GeneratorTaskInputArgs for
layered (PSD) generation output.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Render and manage the separated layers produced by the edit-elements
action as a manual-frame container with attached image and text
children.
- canvasNodeOps: frame container/child helpers, render ordering,
auto-attach nodes to containing frames, and layer reordering
- CanvasWorkspace wires frame-aware selection, drag, resize, hide/lock,
and layer reorder through the new ops
- CanvasPanels: expandable layer tree with rename/reorder
- designGateway.submitGeneratorTask + GeneratorTaskSubmitResponse and
bbox/generator_name artifact metadata in the domain
- ArtboardPreview shows a working state for generating frame layers
- i18n: expandLayer / renameLayer / layerTitlePlaceholder; styles
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Introduce the `layer-separation` generator (edit-elements node action)
that splits a source image into a clean background layer, foreground
image artifacts with bounding boxes, and editable text render data.
- New layer_separation.go builds the separated result and artifacts
- SubmitGeneratorTask routes layer-separation and move-object via a
switch; standalone task completion + polling support
- Extend artifact metadata with BBox and GeneratorName across the
domain, API types, mapper, .api schema, and API.md
- edit-elements gets a transparent manual-frame placeholder holding the
separated foreground images and text layers beside the original
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Switch the editable selection title from onDoubleClick to onClick so
renaming a node/layer title begins with a single click.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Some image endpoints reject the `n` count parameter (Unknown parameter
tools[0].n / param: n). Detect that error and re-issue the batch as
concurrent count=1 requests, each with a suffixed idempotency key, then
assemble the URLs into a single GeneratedImage.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Negative and larger letter spacing values are legitimate for extracted
text layers; clamp only outside ±256 instead of rejecting anything
below 0 (or above 80 in the extractor).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
root