pds

This MCP server provides access to the NASA PDS Registry API via a FastMCP interface, enabling human-in-the-loop agentic search over PDS data products, bundles, collections, and related metadata. It exposes a server-side MVP that can be integrated with MCP clients (such as Claude Desktop, Cursor, or custom clients) to issue search queries, retrieve detailed product metadata, and navigate PDS relationships (missions, targets, instruments, data products). The server leverages the NASA PDS Registry API to resolve URNs, fetch metadata, and present results in a structured way that supports iterative exploration and contextual filtering. To interact with it, configure an MCP client with the server’s mcpServers entry and point the client at the provided Python entry script.

Prerequisites:

• Python 3.13 or newer installed on your system
• git to clone the repository
• network access to install dependencies from requirements.txt

Step-by-step installation:

1. Clone the repository: git clone https://github.com/NASA-PDS/pds-mcp-server.git cd pds-mcp-server

2. Create and activate a virtual environment (recommended): python3.13 -m venv venv source venv/bin/activate

3. Install dependencies: pip install -r requirements.txt

4. Run the server locally to verify: python3.13 pds_mcp_server.py

Note: If you use a different Python environment, ensure the interpreter used to run pds_mcp_server.py matches the one where dependencies were installed.

Environment and usage tips:

• The server is Python-based; ensure Python 3.13+ is used for both installation and runtime.
• If you maintain multiple Python environments, use a virtual environment to isolate dependencies.
• The MCP client configuration example in the README assumes a path to the Python executable and the script; adjust paths according to your deployment.
• For production deployments, consider exposing the MCP server behind a reverse proxy and securing it as needed for your workflow.
• If you encounter issues with the PDS Registry API, verify network access and consider setting appropriate timeouts or retry policies at the client level.