- Use invisible fallback path (M0 0 L0 0) instead of dot for empty d attr
- Track __sourceD on path objects to detect d changes requiring recreation
- Generalize textRecreated to objectRecreated for path recreation support
- Tighten isBadgeLikeFrame: max height 40px, max 3 children, 1 icon max,
exclude explicit button names, width cap 260px, CJK label limit 10 chars
- Use full capsule radius (height/2) instead of clamped range for badges
Add COPYWRITING section to CHAT_SYSTEM_PROMPT, DESIGN_GENERATOR_PROMPT,
and SUB_AGENT_PROMPT enforcing concise text in generated designs:
headlines 2-6 words, descriptions ≤20 words, no 3+ sentence paragraphs.
- Add splitByGrapheme for CJK text to enable natural character-level wrapping
- Add optical vertical centering offset for single-line text in horizontal layouts
- Use fallback path data when SVG d attribute is empty to prevent render errors
- Fix stroke-only path detection to check for visible stroke (non-zero width)
- Add isBadgeLikeFrame heuristic for badge/chip/tag detection
- Enhance badge sizing: auto corner-radius, icon normalization, CJK-aware padding
- Fix fill-style icon resolution to always apply visible fill color
- Add dot/bullet/point/circlefill to icon path map
- Introduce support for .op file format alongside .pen
- Add MCP server functionality for document management and tool operations
- Implement batch processing tools for design and variable management
- Enhance save and open dialogs to accommodate new file format
- Update dependencies and scripts for MCP server compilation and execution
- Pass style guide from orchestrator plan to sub-agent prompts
- Add orchestrator-prompt-optimizer for context-aware prompt tuning
- Improve parallel sub-task generation and error recovery
- Set default lineHeight on text nodes (1.2 for headings, 1.5 for body)
- Expand phone/mockup placeholder detection with shape-based fallback
- Flatten nested phone-shaped frames in post-streaming tree fixes
- Increase CJK character width estimation and button padding buffer
- Update the ORCHESTRATOR_PROMPT to allow for more granular task division, specifying 4-15 atomic sections.
- Enhance the JSON output structure by adding new subtasks for the hero section.
- Adjust ORCHESTRATOR_TIMEOUTS to extend hard and no-text timeout values for improved orchestration performance.
- Remove the complexity classifier as its functionality is no longer needed.
- Update design generator to always route through the orchestrator, simplifying the logic.
- Enhance error handling during orchestration to ensure fallback to direct generation is clear.
- Introduce a new sub-agent prompt for improved output formatting and clarity in design generation.
- Introduce context optimization to manage chat history size and prevent unbounded growth.
- Implement complexity assessment for design prompts to determine orchestration needs.
- Add orchestrator functionality for parallel design generation, improving performance and responsiveness.
- Update chat message parsing to include explicit status handling for orchestrator steps.
- Refactor design generator to support orchestration and streamline context building.
- Add new commands for Electron development, compilation, and building processes in CLAUDE.md.
- Update README.md to reflect the availability of the application as both a web and desktop app.
- Introduce a FixedChecklist component in the AI chat panel for better user interaction with generated tasks.
- Implement CI/CD workflows for automated testing and Electron builds in GitHub Actions.
- Refactor design generator prompts to support element-by-element streaming for improved performance.
- Introduce new icon files in various formats (ICNS, ICO, PNG) for application branding.
- Add logo image for the Electron interface to enhance visual identity.
- Update canvas selection logic to allow programmatic selection from the layer panel.
- Introduce a new ExportSection component to facilitate exporting layers as raster images in various formats (PNG, JPEG, WEBP) with adjustable scale options.
- Integrate the ExportSection into the PropertyPanel for easy access during design editing.
- Enhance export utility functions to support exporting layers and managing descendant nodes.
- Add support for OpenCode as a new AI provider, allowing users to connect to a local OpenCode server.
- Implement new API endpoints for generating and streaming chat responses via OpenCode.
- Update UI components to include OpenCode in provider selection and display.
- Introduce OpenCode logo and enhance the agent settings dialog to manage OpenCode connections.
- Refactor existing chat and generation logic to accommodate provider-specific routing.
- Include type definitions for OpenCode SDK to ensure type safety and improve developer experience.
- Update AI design generation functions to incorporate document variables and themes, allowing for dynamic references in generated designs.
- Modify the buildContextString function to include variable summaries for improved context during AI interactions.
- Enhance prompts to instruct the AI on using variable references instead of hardcoded values, ensuring consistency and flexibility in design outputs.
- Refactor related types and services to support the new variable and theme structures, improving overall design generation capabilities.
- Update history store to support up to 300 undo/redo states, improving user experience during complex editing sessions.
- Refactor batch handling in various components to ensure accurate history tracking and prevent unnecessary entries.
- Implement checks to avoid adding duplicate states to the undo stack, optimizing performance and memory usage.
- Enhance event handling in canvas and keyboard shortcuts to maintain consistent history state during user interactions.
- Increase undo history to 300 states, deduplicate identical snapshots
- endBatch accepts optional currentDoc to skip no-op batches
- Add AI streaming guards: disable canvas interaction during generation
- Defer transform batch close via rAF so object:modified fires before endBatch
- Restrict layout reorder to move drags only (not scale/rotate handles)
- Extend checkDragReparent to work for any parent container, not just root frames
- Wrap design apply/modify in history batches for proper undo support
- Add variable management capabilities to the document store, including setting, removing, and renaming variables.
- Implement a VariablesPanel in the editor layout for user interaction with variables.
- Enhance UI components (e.g., AppearanceSection, FillSection, LayoutSection, StrokeSection) to support variable binding for opacity, fill color, gap, and padding.
- Update code generation services to output CSS variables and React code with variable references.
- Introduce keyboard shortcuts for toggling the variables panel and improve overall user experience.
- Introduce a "Thinking..." indicator in the AI chat panel to improve user experience during processing.
- Update chat message styles for better spacing and alignment.
- Implement batch undo functionality for number input adjustments to streamline user interactions.
- Refactor canvas event handling to ensure accurate history tracking during drag operations.
- Add keyboard shortcuts for agent settings and improve canvas selection handling during history state restoration.
- Path icons: use uniform scaling (preserve aspect ratio instead of squishing),
fillRule 'evenodd' for compound paths with cutouts, transparent fill for
stroke-only icons, strokeUniform to keep stroke width constant
- Text centering: use actual Fabric rendered height (fontSize * lineHeight)
for cross-axis centering instead of declared height
- Hover cursor: use default arrow instead of move/crosshair on elements
- Stream reliability: server sends keep-alive pings every 15s during API TTFT,
forwards thinking_delta events, client resets timeout on ping/thinking chunks
- Remove silent fallback from streamChat to generateCompletion (was causing
double requests when server is unresponsive)
- Add 3-minute timeout to generateCompletion
- AI prompts: instruct model to preserve icon aspect ratios
Implement Figma-style canvas interactions:
- Hover outlines: solid blue on hovered element, dashed on direct children
- Depth-aware selection: single click selects at current depth, double-click
enters containers progressively, Escape exits frame context
- Root frame clipping: depth-0 frames always clip overflowing children
- Drag-out reparenting: dragging a child outside its root frame detaches it
to the overlapping frame or promotes to root level, with undo support
- Layout reorder drag with insertion indicator
- AI design generation with staggered fade-in animation
- Use default cursor on hover instead of move/crosshair
- Separate design generation and modification into dedicated functions
with multi-step iterative pipeline (generateDesign, generateDesignModification)
- Add chat title derived from first message, shown in header and minimized bar
- Add panel resize via top drag handle with height constraints
- Show streaming loader icon next to chat title during AI generation
- Display selected object count in input area for context awareness
- Enhance chat-message with step collapsing and applied state styling
- Expand AI prompts for improved design generation quality
- Add support for 'thinking' state in chat message streaming to indicate processing.
- Update toolbar to parse SVG files into editable nodes, allowing for better integration of SVG graphics.
- Refactor image handling to accommodate both SVG and raster images, ensuring proper placement and scaling on the canvas.
- Improve error handling and user feedback during file uploads.
- Add agent settings dialog to connect local Claude Code and Codex CLI
- Connect button fetches real models from Agent SDK / Codex models cache
- Provider logos (Claude, OpenAI) as React icon components in top-bar and dialog
- Model selector in AI chat panel grouped by provider with logos and Best badge
- MCP CLI integration toggles (Claude Code, Codex, Gemini, OpenCode, Kiro)
- Persist agent settings to localStorage with hydrate on mount
- Redesign AI chat panel: header, empty state pills, input area, color layering
- Add shadcn Switch component, agent-settings types and Zustand store
- new documents start with a 1920x1080 white root frame
- AI-generated content inserts into the root frame instead of document root
- export DEFAULT_FRAME_ID constant for cross-module access
Enhance AI chat and generation functionalities by introducing model selection. Update the chat and generation endpoints to accept an optional model parameter. Implement a new API endpoint to fetch available AI models from the server, and integrate model selection into the AI chat panel for improved user experience.
Implement code generators that convert PenNode trees to React+Tailwind
and HTML+CSS output. Add code panel with tab switching, syntax
highlighting, and copy-to-clipboard. Supports generating code for
selected elements or the entire document.
Add floating AI assistant panel with streaming chat, design keyword
detection, and automatic PenNode JSON extraction to canvas. Server
endpoints use Anthropic SDK (with API key) or Claude Agent SDK (local
OAuth) as fallback. Panel supports minimize, drag-to-snap, and
quick action presets.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
