github-stars
A Cloudflare-powered MCP (Model Context Protocol) Server that allows you to search and query your GitHub starred repositories using natural language.
claude mcp add --transport stdio miantiao-me-github-stars npx -y github-stars
How to use
This MCP server provides a natural language interface to search your GitHub starred repositories. It builds a searchable index from your starred repos, including the repository metadata and processed README content, and serves it through a MCP-compatible API. You can query the data using natural language phrases and receive structured results that highlight relevant repositories and their READMEs. The server is designed to be integrated with AI agents or clients that speak the MCP protocol, enabling seamless querying within larger workflows.
To use the server, first ensure you have deployed and started the MCP service (locally or in Cloudflare Workers). Once running, authenticate your requests with an API key included in the Authorization header: "Authorization: Bearer your-mcp-api-key". You can then call the MCP endpoint (for example, the streamable HTTP URL provided by your deployment) using the MCP tool name search_github_stars and supply a query in natural language, such as: 'Show me the most starred JavaScript projects with README highlights'. The server returns matching repositories along with relevant README content, enabling quick knowledge extraction and exploration of your GitHub stars.
How to install
Prerequisites:
- Node.js v22 or newer
- PNPM package manager
- GitHub Personal Access Token with repo scope
- Cloudflare account (for deployment over Workers) (optional for local development)
Local installation and run:
# Install dependencies
pnpm install
# Fetch your GitHub stars locally
pnpm dev:stars
# Run MCP server locally
pnpm dev:mcp
Deployment to Cloudflare Workers (optional):
pnpm deploy
Notes:
- The workflow expects environmental configuration for Cloudflare R2 and AutoRAG when deploying to Cloudflare. For local development, you primarily need the GitHub token and to install dependencies.
- If you publish and run the server, ensure your MCP_API_KEY is securely stored and supplied in requests to the MCP endpoint.
Additional notes
Tips and common issues:
- Ensure your GitHub token has the repo scope to fetch starred repositories.
- If updates to starred repositories are expected weekly, verify the GitHub Actions workflow and Cloudflare AutoRAG index rebuild triggers are functioning in your deployment.
- For local testing, expose the MCP endpoint or use the streamable HTTP URL provided by your local server setup.
- Set MCP_API_KEY as an environment variable or secret in your deployment environment to protect access to the MCP server.
- If you encounter authentication errors, double-check the Authorization header formatting and token validity.
- The stored data includes repository metadata and processed README content to support richer search results; consider adjusting indexing priorities if your dataset grows large.
Related MCP Servers
supermemory
Your memories are in ChatGPT... But nowhere else. Universal Memory MCP makes your memories available to every single LLM. No logins or paywall. One command to set it up.
mcp-memory
🔥🖥️ MCP Memory is a MCP Server that gives MCP Clients (Cursor, Claude, Windsurf and more) the ability to remember information about users (preferences, behaviors) across conversations.
time
⏰ Time MCP Server: Giving LLMs Time Awareness Capabilities
unity-editor
An MCP server and client for LLMs to interact with Unity Projects
website-publisher
AI Website builder and publisher MCP. Quickly publish and deploy your AI generated code as real website URL. Support html, css, js, python etc.
dubco -npm
The (Unofficial) dubco-mcp-server enables AI assistants to manage Dub.co short links via the Model Context Protocol. It provides three MCP tools: create_link for generating new short URLs, update_link for modifying existing links, and delete_link for removing short links.