MCP Science
Open Source MCP Servers for Scientific Research
What is MCP Science?
MCP Science is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to open source mcp servers for scientific research
Open Source MCP Servers for Scientific Research
This server falls under the Data Science & ML category on MCPgee, the world's largest MCP server directory with 33,000+ servers.
Features
- Open Source MCP Servers for Scientific Research
Use Cases
Maintainer
Works with
Installation
Manual Installation
npx mcp-scienceConfiguration
Configuration Details
claude_desktop_config.json
Performance
Response Metrics
Resource Usage
How to Set Up and Use MCP Science
MCP.science is a collection of open-source MCP servers designed specifically for scientific research workflows, maintained by the Path Integral Institute. It provides a unified toolkit that lets AI assistants like Claude execute Python code in sandboxes, fetch and process HTML and PDF documents, query materials-science databases, run DFT calculations via GPAW, interact with live Jupyter kernels, perform academic and web searches, and more. Researchers can connect any MCP-compatible client to these servers and have AI-driven access to a wide range of scientific data sources and computational tools without writing custom integrations.
Prerequisites
- Python environment with uv installed: curl -sSf https://astral.sh/uv/install.sh | bash
- An MCP-compatible client such as Claude Desktop, VSCode with MCP extension, Goose, or 5ire
- Materials Project API key if using the materials-science server (free at materialsproject.org)
- TXYZ API key if using the academic search server
Install uv
Install the uv package manager, which is used to run mcp-science servers without a separate install step.
curl -sSf https://astral.sh/uv/install.sh | bashChoose the server you need
MCP.science bundles multiple servers: web-fetch (HTML/PDF fetching), python-code-execution (sandboxed Python), materials-project (materials science data), ssh-exec (remote command execution), txyz-search (academic search), jupyter-act (Jupyter kernel interaction), gpaw (DFT calculations), mathematica-check (Wolfram Language evaluation), tinydb (JSON database), timer, and nemad (neuroscience data).
Run a server with uvx
Use uvx to launch any mcp-science server directly without installing it globally. Replace <server-name> with the name of the server you want, such as web-fetch or python-code-execution.
uvx mcp-science web-fetchAdd the server to your MCP client config
Open your Claude Desktop or other MCP client configuration file and add the server entry. The example below adds the web-fetch server.
{
"mcpServers": {
"web-fetch": {
"command": "uvx",
"args": ["mcp-science", "web-fetch"]
}
}
}Add API keys for servers that require them
The Materials Project server and TXYZ search server require API keys. Add them as environment variables in the server config block.
{
"mcpServers": {
"materials-project": {
"command": "uvx",
"args": ["mcp-science", "materials-project"],
"env": {
"MATERIALS_PROJECT_API_KEY": "your_key_here"
}
}
}
}Restart your MCP client and verify
Restart Claude Desktop (or your MCP client) so it picks up the new server configuration. The server name should appear in the connected tools list.
MCP Science Examples
Client configuration
Example Claude Desktop config adding the python-code-execution and web-fetch servers from mcp-science.
{
"mcpServers": {
"mcp-science-python": {
"command": "uvx",
"args": ["mcp-science", "python-code-execution"]
},
"mcp-science-webfetch": {
"command": "uvx",
"args": ["mcp-science", "web-fetch"]
}
}
}Prompts to try
Example prompts that demonstrate the scientific research capabilities of mcp-science servers.
- "Fetch and summarize the abstract from this arxiv PDF: https://arxiv.org/pdf/2301.00001"
- "Run this Python script in a sandbox and return the output: import numpy as np; print(np.linalg.eigvals([[1,2],[3,4]]))"
- "Search for recent academic papers on protein folding using TXYZ search"
- "Look up the crystal structure of Silicon (Si) in the Materials Project database"
- "Start a timer for 25 minutes and notify me when it's done"Troubleshooting MCP Science
uvx command not found after installing uv
Restart your shell or run 'source ~/.bashrc' (or ~/.zshrc) to reload the PATH. The uv installer adds itself to PATH but requires a shell reload.
Materials Project or TXYZ server returns authentication errors
Ensure the correct API key is set in the env block of the server config. Get a free Materials Project API key at next.materialsproject.org/api and a TXYZ key from txyz.ai.
Server name not recognized when running 'uvx mcp-science <name>'
Check the exact server name by visiting the mcp.science GitHub repository for the current list of available server names. Names are case-sensitive and hyphen-separated.
Frequently Asked Questions about MCP Science
What is MCP Science?
MCP Science is a Model Context Protocol (MCP) server that open source mcp servers for scientific research It connects AI assistants to external tools and data sources through a standardized interface.
How do I install MCP Science?
Follow the installation instructions on the MCP Science GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.
Which AI clients work with MCP Science?
MCP Science works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.
Is MCP Science free to use?
Yes, MCP Science is open source and available under the MIT license. You can use it freely in both personal and commercial projects.
MCP Science Alternatives — Similar Data Science & ML Servers
Looking for alternatives to MCP Science? Here are other popular data science & ml servers you can use with Claude, Cursor, and VS Code.
Ultrarag
★ 5.6kA Low-Code MCP Framework for Building Complex and Innovative RAG Pipelines
RocketRide
★ 3.1k📇 🏠 - MCP server that exposes RocketRide AI pipelines as t
Aix Db
★ 2.1kAix-DB 基于 LangChain/LangGraph 框架,结合 MCP Skills 多智能体协作架构,实现自然语言到数据洞察的端到端转换。
NeMo Data Designer
★ 1.9k🎨 NeMo Data Designer: Generate high-quality synthetic data from scratch or from seed data.
PaperBanana
★ 1.7kOpen source implementation and extension of Google Research’s PaperBanana for automated academic figures, diagrams, and research visuals, expanded to new domains like slide generation.
MiniMax
★ 1.5kBridges MiniMax AI capabilities to the Model Context Protocol, enabling AI agents to perform image understanding, text-to-image generation, and speech synthesis. It provides a standardized interface for accessing MiniMax's core tools via JSON-RPC.
Browse More Data Science & ML MCP Servers
Explore all data science & ml servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.
Set Up MCP Science in Your Editor
Choose your AI client for step-by-step setup instructions.
Quick Config Preview
Add this to your claude_desktop_config.json or .cursor/mcp.json
Ready to use MCP Science?
Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.