Neura Market
andybrandt/mcp-simple-pubmed logo

andybrandt/mcp-simple-pubmed

Free

🐍 ☁️ MCP to search and read medical / life sciences papers from PubMed.

FreeFree tier
Type
Open Source
Visit Website GitHub (40 stars)

About andybrandt/mcp-simple-pubmed

MCP Simple PubMed is an open-source MCP server that provides AI assistants with access to PubMed's biomedical literature via the NCBI Entrez API. It offers tools to search PubMed using keywords, MeSH terms, author names, date ranges, and Boolean operators; retrieve full-text articles (when available from PubMed Central); and access abstracts and metadata through resource URIs. The server also includes MCP prompts for constructing effective searches: systematic review search strategies, PICO framework queries, and author publication lookups. Installation is available via pip or Smithery, and configuration requires a user email (and optionally an NCBI API key for higher rate limits). Note that full-text availability is limited to open access articles in PubMed Central, and paywalled papers cannot be accessed.

Key Features

Search PubMed using keywords, MeSH terms, author names, date ranges, and Boolean operators
Retrieve full-text articles when available (open access in PubMed Central)
Access abstracts and metadata via resource URIs
MCP prompts for systematic review, PICO, and author searches
XML-ized full text output optimized for AI consumption
Installable via pip or Smithery for Claude Desktop

Pros & Cons

Pros
  • Provides direct MCP integration for AI assistants like Claude
  • Supports advanced PubMed search syntax (MeSH, Boolean, date filters)
  • Includes pre-built prompt templates for common search patterns
  • Open source and free to use
  • Easy installation via pip or Smithery
  • XML-ized full text gives AI structured document information
Cons
  • Full text limited to open access articles in PubMed Central
  • Paywalled papers cannot be accessed through the tool
  • Sometimes full text may be unavailable even when accessible manually via browser
  • XML output may not be human-readable
  • Rate limited to 3 requests/second without API key

Best For

Systematic literature reviews with comprehensive search strategiesClinical question answering using the PICO frameworkFinding all publications by a specific authorGeneral biomedical and life sciences researchIntegrating PubMed search capabilities into AI-assisted research workflows

FAQ

Why can't I get the full text for some papers?
Full text is only available for open access articles in PubMed Central. Additionally, some papers may appear unavailable through the tool even though they are accessible via a regular browser. The tool cannot access paywalled content.
Does this tool provide access to paywalled papers?
No, the tool cannot access paywalled/paid papers. Users may need institutional library access or alternative methods to read such papers.
What format does the full text come in?
The full text is returned in XML format, which is structured and optimized for AI consumption rather than human reading.
What configuration is required?
You must provide your email address (PUBMED_EMAIL) as required by NCBI. An optional API key (PUBMED_API_KEY) can be set for higher rate limits (10 requests/second instead of 3).