mymlh
Model Context Protocol Server for MyMLH API v4
claude mcp add --transport stdio wei-mymlh-mcp-server docker run -i wei/mymlh-mcp-server \ --env CF_API_TOKEN="your Cloudflare API token with Workers permissions" \ --env MCP_ENDPOINT="https://mymlh-mcp.git.ci/mcp" \ --env CF_ACCOUNT_ID="your Cloudflare account id" \ --env MCLH_CLIENT_ID="your MyMLH OAuth client id" \ --env MCLH_CLIENT_SECRET="your MyMLH OAuth client secret"
How to use
This MCP server provides OAuth-authenticated access to MyMLH data via the Model Context Protocol. After starting the server, you can connect MCP clients using the Streamable HTTP transport with OAuth to the endpoint exposed by the deployment (e.g., the provided Cloudflare Workers edge endpoint). Once connected, you can use the available tools to fetch user data and tokens. The core tools exposed by this server include mymlh_get_user to fetch the current MyMLH user profile, mymlh_get_token to view the current access token details, and mymlh_refresh_token to refresh and persist tokens. These tools allow MCP clients to act on behalf of users with secure, delegated access to their MyMLH data. The server is designed to run at edge via Cloudflare Workers for low latency and scalable access across regions.
How to install
Prerequisites:
- Docker installed and running on your host (recommended for this deployment).
- A Cloudflare account with Workers permissions if you plan to deploy on Cloudflare Workers.
- MyMLH OAuth credentials (client id and client secret).
Install steps:
-
Pull or build the Docker image for the MCP server (assuming the image wei/mymlh-mcp-server is available):
docker pull wei/mymlh-mcp-server
-
Run the MCP server container with required environment variables:
docker run -it --rm
-e CF_ACCOUNT_ID="<your-cloudflare-account-id>"
-e CF_API_TOKEN="<your-cloudflare-api-token>"
-e MCLH_CLIENT_ID="<your-mlh-oauth-client-id>"
-e MCLH_CLIENT_SECRET="<your-mlh-oauth-client-secret>"
-e MCP_ENDPOINT="https://mymlh-mcp.git.ci/mcp"
wei/mymlh-mcp-server -
If using a local development environment or alternative runtimes, ensure the same environment variables are provided and the server exposes the MCP endpoint as documented.
-
Update your MCP clients configuration to point to the deployed endpoint (see the README examples for client snippets).
Additional notes
Notes and tips:
- Ensure the OAuth credentials for MyMLH are securely stored and rotated as needed. The server will refresh tokens automatically if configured.
- When using Cloudflare Workers, consider enabling appropriate CORS and security settings to prevent unauthorized access.
- If you encounter 401 errors, verify that the OAuth flow is supported by your MCP client and that the access token is valid for the MyMLH API scope.
- The server supports a fallback option via mcp-remote for environments that do not support Streamable HTTP with OAuth; see the README for the exact configuration snippet to use with mcp-remote.
- Keep the MCP endpoint URL consistent across all clients to avoid connectivity issues.
Related MCP Servers
context7
Context7 MCP Server -- Up-to-date code documentation for LLMs and AI code editors
obsidian -tools
Add Obsidian integrations like semantic search and custom Templater prompts to Claude or any MCP client.
MiniMax -JS
Official MiniMax Model Context Protocol (MCP) JavaScript implementation that provides seamless integration with MiniMax's powerful AI capabilities including image generation, video generation, text-to-speech, and voice cloning APIs.
1xn-vmcp
vMCP - Virtual Model Context Protocol
mcp-bundler
Is the MCP configuration too complicated? You can easily share your own simplified setup!
akyn-sdk
Turn any data source into an MCP server in 5 minutes. Build AI-agents-ready knowledge bases.