Code Context MCP Server

A Model Context Protocol (MCP) server for providing code context from local git repositories. This server allows you to:

1. Clone git repositories locally
2. Process branches and files
3. Generate embeddings for code chunks
4. Perform semantic search over code

Features

• Uses local git repositories instead of GitHub API
• Stores data in SQLite database
• Splits code into semantic chunks
• Generates embeddings for code chunks using Ollama
• Provides semantic search over code

Prerequisites

• Node.js (v16+)
• Git
• Ollama with an embedding model

Installation

# Clone the repository
git clone <repository-url>
cd code-context-mcp

# Install dependencies
npm install

# Build the project
npm run build

Configuration

Set the following environment variables:

• DATA_DIR: Directory for SQLite database (default: '~/.codeContextMcp/data')
• REPO_CACHE_DIR: Directory for cloned repositories (default: '~/.codeContextMcp/repos')

Using Ollama

For faster and more powerful embeddings, you can use Ollama:

# Install Ollama from https://ollama.ai/

# Pull an embedding model (unclemusclez/jina-embeddings-v2-base-code is recommended)
ollama pull unclemusclez/jina-embeddings-v2-base-code

Usage

Using with Claude Desktop

Add the following configuration to your Claude Desktop configuration file (claude_desktop_config.json):

{
  "mcpServers": {
    "code-context-mcp": {
      "command": "/path/to/your/node",
      "args": ["/path/to/code-context-mcp/dist/index.js"]
    }
  }
}

Tools

The server provides the following tool:

queryRepo

Clones a repository, processes code, and performs semantic search:

{
  "repoUrl": "https://github.com/username/repo.git",
  "branch": "main", // Optional - defaults to repository's default branch
  "query": "Your search query",
  "keywords": ["keyword1", "keyword2"], // Optional - filter results by keywords
  "filePatterns": ["**/*.ts", "src/*.js"], // Optional - filter files by glob patterns
  "excludePatterns": ["**/node_modules/**"], // Optional - exclude files by glob patterns
  "limit": 10 // Optional - number of results to return, default: 10
}

The branch parameter is optional. If not provided, the tool will automatically use the repository's default branch.

The keywords parameter is optional. If provided, the results will be filtered to only include chunks that contain at least one of the specified keywords (case-insensitive matching).

The filePatterns and excludePatterns parameters are optional. They allow you to filter which files are processed and searched using glob patterns (e.g., **/*.ts for all TypeScript files).

Database Schema

The server uses SQLite with the following schema:

• repository: Stores information about repositories
• branch: Stores information about branches
• file: Stores information about files
• branch_file_association: Associates files with branches
• file_chunk: Stores code chunks and their embeddings

Debuggin

curl http://localhost:11434/api/embed -d '{"model":"unclemusclez/jina-embeddings-v2-base-code","input":"Llamas are members of the camelid family"}' curl http://127.0.01:11434/api/embed -d '{"model":"unclemusclez/jina-embeddings-v2-base-code","input":"Llamas are members of the camelid family"}' curl http://[::1]:11434/api/embed -d '{"model":"unclemusclez/jina-embeddings-v2-base-code","input":"Llamas are members of the camelid family"}'

License

MIT