tsilva/simple-wandb-mcp-server

🪄 simple-wandb-mcp-server

A minimal, blazing-fast Model Context Protocol (MCP) server for Weights & Biases (W&B) projects.

✨ Features

• 🚀 Lightweight and easy to deploy
• 🔌 Simple HTTP API for querying W&B runs, metrics, and configs
• 📊 Fetch run details, metrics, and plots in seconds
• 🛠️ Built for integration with LLMs, dashboards, and automation
• 🐍 Pure Python, no heavy dependencies

⚙️ Setup

# Clone the repository
$ git clone https://github.com/yourusername/simple-wandb-mcp-server.git
$ cd simple-wandb-mcp-server

# (Optional) Create and activate a virtual environment
$ python -m venv venv
$ source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install dependencies
$ pip install -r requirements.txt  # or pip install .

🚀 Usage

Start the server:

$ python server.py

The server will start on http://localhost:8000 by default.

Example API Calls

• Get W&B run details:
  • GET /wandb/run?entity=<entity>&project_name=<project>&run_id=<run_id>
• List all runs in a project:
  • GET /wandb/runs?entity=<entity>&project_name=<project>
• List all metrics in a project:
  • GET /wandb/metrics?entity=<entity>&project_name=<project>
• Plot a metric for a run:
  • GET /wandb/plot?entity=<entity>&project_name=<project>&run_id=<run_id>&metric_names=<metric1,metric2>

🧩 Integrations

• Works out-of-the-box with Weights & Biases
• Designed for easy extension and embedding in larger systems

📄 License

This project is licensed under the MIT License.

About

A simple, production-ready MCP server for W&B. Perfect for LLM agents, dashboards, and automation scripts that need fast access to experiment metadata and metrics.

Contributing

Pull requests and issues are welcome! Feel free to open an issue or submit a PR.

Author

• tsilva