MCP.science Servers

MCPOpen SourceMIT62.0

by pathintegral-institute • Data & Databases

A collection of open-source MCP servers that let AI models interact with scientific data, tools, and workflows.

Example Use Cases

1
Fetch and summarise HTML, PDF, and plain-text web content for evidence-backed answers.
2
Execute sandboxed Python code or run computations (including GPAW DFT) as part of analysis pipelines.
3
Query scientific databases (e.g., Materials Project), interact with Jupyter kernels, or run pre-approved SSH commands to access remote tools.

Description

This repository provides multiple MCP (Model Context Protocol) servers tailored for scientific research, enabling LLMs to access web content, run sandboxed Python, query materials databases, control remote machines, execute DFT computations, interact with Jupyter kernels, and more. Each server is packaged for easy launch (via uvx) and follows a standardized protocol so AI clients can plug into diverse data sources and tools. The project is MIT-licensed and intended to accelerate reproducible, secure integrations between models and scientific infrastructure.

Capabilities(2 total)

Tools (2)

fetch-web

Fetch a URL and return the content. Images will be returned in ![]() format. DO NOT FETCH image_urls, for images use read-image-url instead.

url:string*

read-image-url

Read images from URLs, convert to LLM readable format.

image_urls:array*

Quick Actions

View on GitHub

Security

Scanned 3 month(s) ago

Risk Level

MINIMAL

Read-only data retrieval, no side effects

Trust Score

C58/100

5/17 checks passed

Scores are informational only and provided “as is” without warranty. AgentHotspot assumes no liability for actions taken based on these ratings.

Quick Stats

Service TypeMCP

Pricing ModelFree

Capabilities2 Tools / 0 Prompts / 0 Resources

Ownerpathintegral-institute

CategoryData & Databases

DependenciesStandalone

Set Your Username