github-to

GitHub to MCP generates an MCP server from a GitHub repository, exposing a set of tools that let AI assistants read code, explore the repo structure, and interact with APIs or documentation present in the project. The generated server includes tooling derived from the repository, such as file readers, code search utilities, OpenAPI/GraphQL endpoints (when discovered), and executable CLI or HTTP routes described by the repo. You can run the server via the recommended CLI command (npx @nirholas/github-to-mcp <repo-url>) and then download the resulting MCP package or use the provided configuration to connect to Claude Desktop, Cursor, Windsurf, or other MCP clients. The web UI also offers an interactive playground to test the tools before integration.

Prerequisites:\n- Node.js (14+ or 16+) and npm/yarn/pnpm installed on your machine.\n- Access to a terminal or shell.\n\nInstallation steps:\n1) Install and run via npx (no local install required):\n\n npx -y @nirholas/github-to-mcp https://github.com/owner/repo\n\n2) (Alternative) Install locally and run a CLI server from the generated package:\n\n npm install -g @nirholas/github-to-mcp\n github-to-mcp https://github.com/owner/repo\n\n3) If you want to build a standalone MCP server package from the repo for deployment, use the web interface or the CLI to generate the server, then deploy the produced package according to your environment (e.g., Node environment, Docker, or Python if available).\n\nPrerequisites overview:\n- A GitHub repository URL to convert (public or private with a token).\n- Access to the internet for dependency installation.\n- Basic familiarity with running CLI commands and deploying MCP packages.

Tips and common issues:\n- For private repos, provide a GitHub token to increase rate limits: GITHUB_TOKEN=your_token_value.\n- The generated MCP server may rely on repository structure (OpenAPI specs, GraphQL schemas, and documented CLI commands). If tools aren’t appearing as expected, ensure the repository contains recognizable APIs or CLI interfaces and that extraction options in the web UI are configured appropriately.\n- When deploying, consider using the provided Docker templates or deployment scripts if available in the generated package.\n- If the repository uses authentication for API calls (e.g., private APIs), ensure any required environment variables or secrets are supplied in your deployment environment.\n- Update frequency depends on the generator’s support for the target language (TypeScript or Python). For production, pin the generated MCP server version to avoid unexpected breaking changes.