Figma MCP Server

What is Figma MCP Server?

A Model Context Protocol (MCP) server that provides read‑only access to Figma files and projects through Claude and other MCP‑compatible clients. It is designed for designers and developers who want to query Figma design data interactively, with a server architecture that also supports future design token and theme management features (pending Figma API enhancements or plugin development).

How to use Figma MCP Server?

Install the npm package, create a .env file with your FIGMA_ACCESS_TOKEN and optionally set MCP_SERVER_PORT and DEBUG flags, then start the server. For Claude Desktop, add an entry in the claude_desktop_config.json using an absolute path to the compiled dist/index.js. Invoke tools like get-file, list-files, create-variables, or create-theme through MCP client requests. (Note: write operations are limited by Figma API restrictions – personal access tokens only support read operations.)

Key features of Figma MCP Server

• Secure authentication with Figma API
• Read file details and list project files
• Performance optimised with LRU caching and rate‑limit handling
• Health checks, usage statistics, and error tracking
• Both stdio and SSE transport protocols supported
• Designed for future variable, theme, and reference management

Use cases of Figma MCP Server

• Querying Figma file structure and component details from an AI assistant
• Listing all files in a Figma project to find relevant designs
• Automating design system audits by retrieving variable and theme data (when API permits)
• Integrating Figma data into custom workflows without leaving the MCP ecosystem

FAQ from Figma MCP Server

What API restrictions apply?

Personal access tokens support only read operations. Write tools (create‑variables, create‑theme) are included in the code but cannot modify Figma files through the REST API. Full write functionality would require Figma plugin development or expanded API access.

What runtime and dependencies are required?

Node.js 18.x or higher, a Figma access token with appropriate permissions, and basic knowledge of the Model Context Protocol.

How does the server handle caching and rate limits?

An LRU cache with a default 5‑minute TTL and a maximum of 500 entries is used. The server follows Figma API rate limits and implements exponential backoff for better handling.

Which transports and authentication are supported?

Both stdio and SSE transport mechanisms are available. Authentication is limited to personal access tokens; OAuth support is planned for the future. No team‑level permissions or collaborative editing are supported.

Where are configuration and log files located?

For Claude Desktop: on macOS at ~/Library/Application Support/Claude/claude_desktop_config.json, on Windows at %APPDATA%\Claude\claude_desktop_config.json. Server logs are at ~/Library/Logs/Claude/mcp*.log (macOS) or %APPDATA%\Claude\logs\mcp*.log (Windows).