Development · Browser extension
pen.dev
Agentic canvas and IDE extension for building UI designs directly into code
pen.dev (formerly Pencil) is an open-format design canvas and IDE extension built for human-AI software design. It uses an open JSON-based .pen file format and connects to coding agents like Claude Code or Cursor via MCP to generate, visually tweak, and export production UI.
What it's for
- Design and iterate UI visually inside VS Code or Cursor using AI coding agents
- Connect AI agents via MCP to inspect and edit design files directly in a repository
- Export visual frames directly to HTML, CSS, or Tailwind markup
- Import existing live websites or Figma files into editable canvas layers
Features
- Open .pen file format structured as agent-readable JSON
- MCP server and CLI support for integration with Claude Code, Cursor, and Windsurf
- WebGL-powered infinite canvas with layers, components, and layout tools
- Built-in browser and Chrome extension for importing live HTML pages as layers
- Direct Figma and .fig file import support
- One-click export to HTML and Tailwind CSS
- Native visual shader effects and vector pen tool
Pricing
Freemium. Free tier available; paid plans for extended team and usage features.
Details
- Type
- Browser extension
- Category
- Development
- Topic
- UI Design & Prototyping
- Open source
- No
- Platforms
- macOS, Windows, Command line, Browser extension
- For
- Designers and developers building software UIs alongside AI coding agents
- Alternative to
- Figma, Sketch, v0
- Tags
- ui design, mcp, vscode extension, ai agents, cursor, tailwind
Why it was saved
- @casey.aicreates: Solves the generic or messy look of AI-generated 'vibe coded' UIs by providing design templates, a visual canvas editor, and automatic MCP integration for AI agents to build and export polished UI components.
- @LinusEkenstam: Generative UI design in an infinite canvas with design-system components, style guides, and direct export to code via Claude Code (MCP).
Compare
Related
Design vocabulary, rules, and skills for AI coding agents.
Agent-first control plane backed by 2,400+ curated AI developer skills
React and Tailwind dashboard UI kit with agentic progress and chat components.
Agent harness optimization system with skills, hooks, and configs for coding agents
Pattern for using Model Context Protocol to trigger Zapier multi-app automations
Walkthrough for setting up Model Context Protocol servers inside Cursor
Updated