NASA

This NASA MCP Server exposes a collection of NASA public APIs through MCP endpoints, enabling you to fetch astronomy imagery and data such as APOD, Mars rover images, NEO feeds, EPIC Earth images, and GIBS satellite imagery, along with an image analysis tool. You can call each tool by its name to retrieve data, images, or analysis-ready content. The server is designed to be used with MCP clients configured to point at the nasa-mcp-server command via uvx, with your NASA API key supplied in the environment variable NASA_API_KEY. The included tools provide structured parameters (dates, dates ranges, sol values, camera options, bounding boxes, and more) to tailor requests to NASA’s services.

Prerequisites:

• Python installed (for uvx-based MCP usage for this server)
• Access to NASA API key (register at https://api.nasa.gov/)
• MCP client tooling (uvx) installed

Installation steps:

1. Install the MCP server package from PyPI: pip install nasa-mcp-server

2. Run the MCP server using uvx (recommended approach): uvx nasa-mcp-server

3. Set up your NASA API key in your environment before starting the server. Example: export NASA_API_KEY=YOUR_NASA_API_KEY_HERE uvx nasa-mcp-server

4. (Optional) Configure client-side mcp.json (VS Code or Claude Desktop) to point to the nasa-mcp-server with the NASA_API_KEY in the environment, as shown in the README examples.

Tips and caveats:

• Ensure NASA_API_KEY is set in the environment where the MCP server runs; without it, requests to NASA services may fail.
• The available tools cover APOD, Mars rover imagery, NEO data, EPIC Earth imagery, and GIBS imagery. Each tool accepts specific parameters (dates, ranges, camera types, bounding boxes, etc.); consult the README for valid argument names.
• For large image requests (GIBS) consider adjusting width, height, and bbox to manage bandwidth.
• If you encounter API rate limits, consider spacing requests or batching where supported.
• The Image Analysis Tool (get_image_analyze) can process images from any URL and returns a base64-encoded payload suitable for LLM ingestion; ensure the URL is publicly accessible.
• If running behind a firewall or proxy, ensure outbound access to NASA API endpoints is permitted and that uvx is able to set environment variables correctly.