home / mcp / fast context mcp server

Fast Context MCP Server

AI-driven semantic code search for codebases via a local MCP server with automatic Windsurf API integration.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "sammysnake-d-fast-context-mcp": {
      "command": "node",
      "args": [
        "/absolute/path/to/fast-context-mcp/src/server.mjs"
      ],
      "env": {
        "WINDSURF_API_KEY": "YOUR_WIND SURF API KEY"
      }
    }
  }
}

You can run Fast Context MCP as a local server that lets any MCP-compatible client search codebases using natural language. It orchestrates local tooling to map your project, run AI-guided searches, and return precise file paths and line ranges with helpful keywords.

How to use

Launch the local MCP server and connect your MCP client (Claude Code, Claude Desktop, Cursor, or any MCP-compatible client). The server maps your project to a virtual codebase, asks Windsurf’s AI for search guidance, and then automatically runs fast, local commands to locate relevant code. You’ll receive a list of relevant files, targeted line ranges, and suggested search keywords to refine future queries.

How to install

Prerequisites: ensure you have Node.js version 18 or higher installed on your system.

Clone the repository, install dependencies, and prepare to run the MCP server locally.

Installation commands

git clone https://github.com/SammySnake-d/fast-context-mcp.git
cd fast-context-mcp
npm install

Setup and configuration notes

The server can auto-detect your Windsurf API key from your local Windsurf installation. If you prefer, you can provide the key directly by setting the environment variable WINDSURF_API_KEY when starting the server.

Additional configuration and usage tips

You can run the provided MCP tool to extract the Windsurf API key from your installation if you do not supply it manually. After setup, you can configure your MCP client to point at the local server binary and pass the API key in the environment section.

Troubleshooting and tips

If you encounter errors related to payload size or timeouts, try adjusting the maximum search rounds and command concurrency in your client configuration. The server documents diagnostic hints to help you tune parameters.

Notes

The server bundles necessary tools (ripgrep and a directory tree utility) so you do not need system-wide dependencies. This enables cross-platform usage on macOS, Windows, and Linux.

Security considerations

Keep your Windsurf API key secure. Do not expose it in shared configurations. Use per-user or per-environment configurations where possible.

Availability of tools and integration

The MCP server exposes a semantic search tool that accepts a natural language query and returns relevant files, suggested keywords, and diagnostic metadata to help you fine-tune searches.

Example workflow

1) Start the local MCP server. 2) In your MCP client, select the Fast Context MCP server as the search backend. 3) Enter a natural language query like, ”where is the authentication logic?” 4) Review the list of relevant files and suggested keywords, then refine your query or open the files directly.

Available tools

fast_context_search

AI-driven semantic code search with tunable parameters to locate relevant code sections across a project.

extract_windsurf_key

Extract Windsurf API Key from the local Windsurf installation for automatic server configuration.