figma-mcp-proxy

# Figma MCP Proxy A Go-based HTTP proxy server that enhances Model Context Protocol (MCP) tool calls by automatically adding Figma file parameters and opening Figma designs when processing tool calls. - For more information, check out the [Cascading: Cloud AI Implements Figma and Jira Usage Guide](https://bitovi.atlassian.net/wiki/spaces/AIEnabledDevelopment/pages/1517289538/Cascading+v2+Cloud+AI+implements+Figma+and+Jira). - Need help? Find Bitovi on [Discord](https://discord.gg/J7ejFsZnJ4) or [hire us](https://www.bitovi.com/services/ai-consulting). 👉 Bitovi can help you integrate this into your own SDLC workflow: [AI for Software Teams](https://www.bitovi.com/ai-for-software-teams) ## What it does This proxy acts as an intermediary between MCP clients and MCP servers, providing the following functionality: ### 1. Tool Schema Enhancement When a client requests the list of available tools (`tools/list`), the proxy: - Identifies tools that have a `nodeId` parameter in their input schema - Automatically adds two additional parameters to these tools: - `fileKey`: The Figma file identifier (extracted from URLs like `https://figma.com/design/1234/5678?node-id=1-2`) - `fileName`: The Figma file name (extracted from the same URL format) - Updates tool descriptions to explain how to extract these parameters from Figma URLs ### 2. Automatic Figma Design Opening When processing tool calls that include both `fileKey` and `fileName` parameters, the proxy: - Automatically opens the specified Figma design using the system's default Figma application - Uses the `figma://` URL scheme to launch directly to the design - Supports macOS, Windows, and Linux operating systems - Adds a 2-second delay to allow Figma to fully launch before proceeding ## Configuration The proxy can be configured using environment variables: - `TARGET_URL`: The MCP server to proxy requests to (default: `http://localhost:3845`) - `PORT`: The port to run the proxy server on (default: `3846`) ## Usage ### Starting the proxy ```bash go run main.go ``` Or with custom configuration: ```bash TARGET_URL=http://localhost:3000 PORT=8080 go run main.go ```