TuringMind

You set up a type-safe MCP server that lets Claude interact with your cloud services through clearly defined tools. It provides a structured, error-checked flow for authenticating, uploading code reviews, fetching repository context, and submitting feedback, reducing misformatted data and failed requests.

How to use

Use the MCP server by pairing it with your MCP client (such as Claude) to access a predefined set of tools. You can initiate login, complete authentication, validate your session, upload code reviews, retrieve repository context, and submit feedback on issues. The flow is designed to guide you through device code login, automatic API key storage, and clear status messages for each step.

How to install

pip install turingmind-mcp

pipx install turingmind-mcp

git clone https://github.com/turingmindai/turingmind-mcp.git
cd turingmind-mcp
pip install -e .

turingmind-mcp --help

Configuration and usage notes

Configure Claude Desktop to use the MCP server by adding a server entry that points to the turingmind-mcp executable. You can embed the runtime command in the Claude desktop config so the client launches the MCP bridge directly.

{
  "mcpServers": {
    "turingmind": {
      "command": "turingmind-mcp"
    }
  }
}

Login flow and authentication

Start authentication from your MCP client. You’ll receive a verification URL, a user code, and a device code. Open the URL in your browser, enter the user code, and complete sign-in. The client will poll for completion and then save your API key to a local configuration file.

If your environment requires a custom API URL, you can pass it via environment variables to the MCP bridge.

Available actions you can perform through the MCP server

Authenticate and manage access with a device code flow that stores your API key locally.

Upload code reviews to the cloud with structured issue data and a full review content payload.

Fetch repository context to aid your reviews with memory of recent issues and conventions.

Submit feedback on specific issues to mark them as fixed, dismissed, or false positives.

Troubleshooting

If you see a message about API key configuration, complete the login flow or manually set the API key in your environment.

If you encounter a permission error, re-run the login flow to obtain a key with the required permissions.

For connection issues, verify the API URL and network connectivity. When developing locally, ensure the backend is running and reachable.

Available tools

turingmind_initiate_login

Start device code authentication flow. No API key required.

turingmind_poll_login

Poll for authentication completion and save the API key locally.

turingmind_validate_auth

Validate the API key and verify account status.

turingmind_upload_review

Upload code review results to the TuringMind cloud with repository context and issues.

turingmind_get_context

Retrieve memory context for a given repository to inform future reviews.

turingmind_submit_feedback

Submit feedback on a code review issue, marking it as fixed, dismissed, or false positive.