MCP.science: Open Source MCP Servers for Scientific Research 🔍📚

What is MCP.science?

MCP.science is an open-source collection of MCP (Model Context Protocol) servers designed for scientific research. It enables AI models like Claude to interact with scientific data, tools, and resources through a standardized protocol.

How to use MCP.science?

Install the uv tool and an MCP-enabled client (e.g., Claude Desktop). Launch any server with uvx mcp-science <server-name>. Configure the command in your client’s settings (e.g., JSON config for Claude Desktop). A step-by-step guide is available in the repository.

Key features of MCP.science

• Open source MCP servers for scientific research
• Easy launch with uvx mcp-science <server-name>
• Includes servers for materials, web, Python, SSH, and more
• Supports multiple MCP‑enabled clients (Claude Desktop, VSCode, Goose, 5ire)
• Active community contributions under MIT license

Use cases of MCP.science

• Search and visualize materials‑science data from the Materials Project database
• Execute Python code in a sandboxed environment for safe analysis
• Fetch and summarise web content (HTML, PDF, plain text)
• Run pre‑validated commands on remote machines over SSH
• Perform density‑functional‑theory calculations with GPAW

FAQ from MCP.science

What is MCP?

MCP (Model Context Protocol) is an open standard that connects AI models to data sources and tools, acting like a USB-C port for AI applications. It provides a growing list of pre‑built integrations and best practices for data security.

How do I install and launch a server?

First install uv from astral.sh/uv. Then run uvx mcp-science <server-name> in your terminal. The command downloads and executes the server automatically.

Do any servers require API keys?

Yes. The Materials Project server and the TXYZ Search server each require a specific API key. Other servers (e.g., Web Fetch, Python Code Execution) work without one.

What clients are supported?

MCP.science works with any MCP‑enabled client, including Claude Desktop, VSCode, Goose, and 5ire. Configure the server command in the client’s settings.

How can I contribute to this project?

Fork the repository, create a feature branch,