Fast Context MCP

To install from the source repository:

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

You can install the latest stable release or the beta/next release using npm. This project has no system-level dependencies as ripgrep and tree are bundled.

Latest stable release:

npm install @sammysnake/fast-context-mcp

Beta/next release:

npm install @sammysnake/fast-context-mcp@next

Add the server configuration to your claude_desktop_config.json file under the mcpServers key. You must provide your WINDSURF_API_KEY in the env object.

Stable release configuration:

{
  "fast-context": {
    "command": "npx",
    "args": ["-y", "--prefer-online", "@sammysnake/fast-context-mcp"],
    "env": {
      "WINDSURF_API_KEY": "sk-ws-01-xxxxx"
    }
  }
}

Beta/next release configuration:

{
  "fast-context": {
    "command": "npx",
    "args": ["-y", "--prefer-online", "@sammysnake/fast-context-mcp@next"],
    "env": {
      "WINDSURF_API_KEY": "sk-ws-01-xxxxx"
    }
  }
}

Add the server configuration to your ~/.claude.json file under the mcpServers key. You must provide your WINDSURF_API_KEY in the env object.

Stable release configuration:

{
  "fast-context": {
    "command": "npx",
    "args": ["-y", "--prefer-online", "@sammysnake/fast-context-mcp"],
    "env": {
      "WINDSURF_API_KEY": "sk-ws-01-xxxxx"
    }
  }
}

Beta/next release configuration:

{
  "fast-context": {
    "command": "npx",
    "args": ["-y", "--prefer-online", "@sammysnake/fast-context-mcp@next"],
    "env": {
      "WINDSURF_API_KEY": "sk-ws-01-xxxxx"
    }
  }
}

The server is a Node.js application that communicates via the Model Context Protocol (MCP) using standard input/output (stdio) transport. You can start it directly using node.

node src/server.mjs

Attempts to extract the Windsurf API Key from your local installation (Devin, Deviv, or Windsurf) by reading local SQLite databases. This tool takes no parameters.

Note: You can hide this tool from the MCP tool list by setting the environment variable FC_HIDE_EXTRACT_WINDSURF_KEY_TOOL=1 at startup.

The server behavior can be tuned using the following environment variables:

You can control the truncation behavior of the ToolExecutor output by setting the following environment variables. This is useful if you need to increase the amount of context returned to an AI agent.

The ToolExecutor class is designed to execute restricted filesystem and search commands (like rg, ls, readfile, tree, and glob) within a specific project root. It is optimized to mimic Windsurf's behavior, including path remapping (converting real paths to /codebase) and output truncation to prevent overwhelming the context.

To use it, instantiate the class with your project's absolute path. You can then either call specific methods directly or use the execCommand / execToolCall methods to dispatch commands from structured objects.

import { ToolExecutor } from './src/executor.mjs';

const executor = new ToolExecutor('/absolute/path/to/project');

// Direct method call
const files = executor.ls('/codebase', true, false);

// Dispatching via command object
const result = executor.execCommand({
  type: 'rg',
  pattern: 'TODO',
  path: '/codebase'
});

Use extractKeyInfo() to programmatically retrieve information about the discovered Windsurf API key. This is useful for verifying if the user is correctly logged into Windsurf before starting a search.

import { extractKeyInfo } from '@sammysnake/fast-context-mcp';

try {
  const keyInfo = await extractKeyInfo();
  console.log('API Key found:', keyInfo.api_key);
} catch (err) {
  console.error('Could not find API key:', err.message);
}