grok-faf-mcp provides a URL-based MCP interface for Grok using the FAF format. It serves 21 tools that expose AI context management, project initialization, RAG-based retrieval, and various FAF utilities via HTTP-SSE. Point your MCP client to the server's /sse endpoint (for example, https://grok-faf-mcp.vercel.app/sse) to begin streaming tool results and context. The server is designed for zero-configuration deployment: you can run it locally with npx grok-faf-mcp or deploy it to a hosting platform and have instant access to all 21 tools without additional setup. Typical use cases include initializing .faf projects, evaluating AI-readiness, syncing context, and performing quick FAF-guided operations across your workflow.

Prerequisites:

• Node.js (and npm) installed on your machine
• Internet access to fetch the MCP package via npx

Installation steps:

1. Open a terminal.
2. Ensure you have Node.js and npm installed:
   • node --version
   • npm --version
3. Run the MCP server using npx (no global install required): npx grok-faf-mcp

   or to run in the background or in scripts, you can reference the package directly:

   npx -y grok-faf-mcp

4. If you prefer to integrate into a custom MCP config, you can use the example: { "mcpServers": { "grok-faf": { "command": "npx", "args": ["-y", "grok-faf-mcp"] } } }

Notes:

• The server is URL-based and exposes 21 tools via HTTP-SSE. You can run it locally or deploy to Vercel or similar hosting to get a public endpoint.

Tips and caveats:

• The MCP model expects the /sse endpoint to stream tool results. Use a client that supports HTTP-SSE.
• The 21 tools include faf_init, faf_score, faf_about, faf_read, rag_query and more—each provides FAF-specific capabilities for managing project context and AI readiness.
• If you encounter network issues, ensure your hosting environment allows long-lived SSE connections.
• When configuring via MCP, you can reuse the grok-faf endpoint pattern and swap the base URL for your own deployment.
• npm_package name is grok-faf-mcp; consider pinning a version in environments where stability is critical.