mcp-figma
Bridges AI clients to Figma Desktop via Plugin API and WebSocket, enabling real-time design manipulation without rate limits.
README
mcp-figma
MCP server that bridges AI clients (Claude, Cursor, GPT) to Figma Desktop via Plugin API + WebSocket. Zero rate limits, free, real-time design manipulation.
Why This Exists
Figma's REST API is rate-limited (6 calls/month on free tier) and existing MCP servers all wrap that same API. mcp-figma takes a different approach — it uses the Figma Plugin API running inside your desktop app, which has no rate limits and is completely free.
How It Works
Three components run simultaneously:
| Component | What it does | How to start |
|---|---|---|
| WebSocket Server | Relay with token auth + channel routing | npm run ws |
| Figma Plugin | Executes commands inside Figma Desktop | Import in Figma |
| MCP Server | Exposes 25 design tools to AI via stdio | Configure in AI client |
Quick Start
Prerequisites
Step 1: Install & Build
git clone https://github.com/tranhoangtu-it/mcp-figma.git
cd mcp-figma
npm install
npm run build
Step 2: Start WebSocket Server
npm run ws
Output:
[ws] WebSocket server listening on 127.0.0.1:3055
[ws] Session token: abc123... ← Copy this token!
Step 3: Import Figma Plugin
- Open Figma Desktop
- Go to Plugins → Development → Import plugin from manifest...
- Navigate to
mcp-figma/src/figma-plugin/and selectmanifest.json - Run the plugin: Plugins → Development → MCP Figma Bridge
- Paste the session token from Step 2 → Click Connect
- Status dot turns green = connected
Step 4: Configure Your AI Client
Claude Code
Add to your project's .mcp.json:
{
"mcpServers": {
"mcp-figma": {
"command": "node",
"args": ["/absolute/path/to/mcp-figma/dist/mcp-server/index.js"],
"env": {
"MCP_FIGMA_TOKEN": "paste-token-from-step-2"
}
}
}
}
Cursor
Add to MCP settings (Settings → MCP Servers):
{
"mcp-figma": {
"command": "node",
"args": ["/absolute/path/to/mcp-figma/dist/mcp-server/index.js"],
"env": {
"MCP_FIGMA_TOKEN": "paste-token-from-step-2"
}
}
}
Windsurf / Other MCP Clients
Any MCP-compatible client works — configure it to run node dist/mcp-server/index.js with the MCP_FIGMA_TOKEN environment variable.
Step 5: Try It!
Ask your AI: "Create a login page in Figma with email input, password input, and a blue submit button"
Watch it appear in real-time on your Figma canvas.
Available Tools (25)
Read Design
| Tool | Description |
|---|---|
get_document_info |
Document name, pages, current page |
get_selection |
Currently selected nodes |
get_node_info |
Detailed properties of a node by ID |
get_page_nodes |
All top-level nodes on a page |
scan_text_nodes |
Find all text nodes in subtree |
scan_nodes_by_type |
Find nodes by type (FRAME, TEXT, etc.) |
Create Nodes
| Tool | Description |
|---|---|
create_frame |
Frame with position, size, fill, auto-layout |
create_rectangle |
Rectangle with fill and corner radius |
create_text |
Text node (Inter font default) |
create_ellipse |
Ellipse or circle |
Modify Nodes
| Tool | Description |
|---|---|
move_node |
Move to new X/Y coordinates |
resize_node |
Change width and height |
set_name |
Rename a node |
set_corner_radius |
Round corners |
delete_node |
Remove from canvas |
clone_node |
Duplicate a node |
group_nodes |
Group multiple nodes |
set_auto_layout |
Configure auto-layout direction, spacing, padding |
Style Nodes
| Tool | Description |
|---|---|
set_fill_color |
Solid fill color (RGBA 0-1) |
set_stroke |
Stroke color and width |
set_text_content |
Update text content |
set_font_size |
Change font size |
set_opacity |
Set transparency (0-1) |
get_local_styles |
List all local paint/text styles |
Export
| Tool | Description |
|---|---|
export_node |
Export as PNG, SVG, or PDF (base64) |
Configuration
| Environment Variable | Default | Description |
|---|---|---|
MCP_FIGMA_TOKEN |
(required) | Session token from WebSocket server |
MCP_FIGMA_PORT |
3055 |
WebSocket server port |
MCP_FIGMA_CHANNEL |
mcp-figma |
Channel name for communication |
Development
npm run build # Build with tsup
npm run dev # Watch mode (auto-rebuild)
npm run ws # Start WebSocket relay server
npm run start # Start MCP server (needs MCP_FIGMA_TOKEN)
npm test # Run all tests
npm run test:ws # Run WebSocket tests only
Security
| Measure | Details |
|---|---|
| Localhost only | WebSocket binds to 127.0.0.1 — no remote access |
| Token auth | Per-session random token, timing-safe comparison |
| Channel isolation | Clients can only message their own channel |
| Schema validation | All messages validated with Zod before processing |
| Input sanitization | Design data sanitized before reaching AI |
| Rate limiting | 100 messages/sec per client |
| Heartbeat | Dead connections detected and cleaned up |
Troubleshooting
| Problem | Solution |
|---|---|
| Plugin won't connect | Check token matches, WS server is running, port 3055 not blocked |
| "Font not found" error | Figma defaults to Inter — ensure it's installed or specify available font |
| MCP server exits immediately | Check MCP_FIGMA_TOKEN is set and WS server is running |
| Tools timeout after 30s | Ensure Figma plugin is connected (green status dot) |
| Port 3055 in use | Set MCP_FIGMA_PORT=3056 for both WS server and MCP client config |
Contributing
See CONTRIBUTING.md for development setup, conventions, and how to add new tools.
License
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.