MCP Simple Pubmed logo

MCP Simple Pubmed

by andybrandt

Searches PubMed, retrieves abstracts and metadata, and downloads available open-access full text through the Entrez API.

Install MCP Simple Pubmed

Pick your client, then add it in one click or paste the config.

STDIOPython
~/Library/Application Support/Claude/claude_desktop_config.json
claude_desktop_config.json
{
  "mcpServers": {
    "andybrandt-mcp-simple-pubmed": {
      "command": "uvx",
      "args": [
        "mcp-simple-pubmed"
      ],
      "env": {
        "PUBMED_EMAIL": "YOUR_VALUE_HERE",
        "PUBMED_API_KEY": "YOUR_VALUE_HERE"
      }
    }
  }
}

Settings → Developer → Edit Config, merge this into mcpServers, then fully quit and reopen Claude Desktop.

Values are written into the snippet above so you can paste it as is. They never leave your browser.

Install not yet sandbox-tested— we haven't run this command through our automated sandbox yet. Not a red flag; we're working through the catalog.

Overview

andybrandt/mcp-simple-pubmed MCP server connects AI assistants to PubMed through the NCBI Entrez API. It searches with keywords, MeSH terms, author names, date ranges, and Boolean operators, then retrieves abstracts, metadata, and available full text. Its prompts help construct systematic review, PICO, and author-focused searches. Reach for it when an agent needs structured access to medical and life sciences literature rather than general web search.

Use cases

  • Search PubMed for studies using MeSH terms and Boolean operators
  • Retrieve abstracts and metadata for selected articles
  • Download XML full text from available PubMed Central articles
  • Construct systematic review searches with guided prompts
  • Build PICO-based clinical literature queries

Key features

  • PubMed search through the NCBI Entrez API
  • Keyword, MeSH, author, date, and Boolean query support
  • Abstract and article metadata access
  • Available PubMed Central full-text retrieval
  • Systematic review, PICO, and author search prompts

Security & health

What it can reach and how reliably it runs — from our own checks, not the maintainer’s claims.

ActiveLast checked 1d ago

Recent health check succeeded.

22/22 checks healthy over the last 85d

Supply-chain signal

No high-severity advisories surfaced by our automated scan.

Critical 0High 0Medium 0Low 0

Scanned 8d ago via OSV.dev · mcp-simple-pubmed (PyPI)

Adoption & maintenance

Factual signals from GitHub, npm, and our automated checks — not a rating.

GitHub stars171Stargazers on the source repository.
Last commit7mo agoMost recent push to the default branch.
Directory activity6 viewsConfig copies, upvotes, and views on AllMCPs.

How MCP Simple Pubmed works

What andybrandt/mcp-simple-pubmed MCP server does

andybrandt/mcp-simple-pubmed MCP server provides an MCP interface to PubMed articles through the NCBI Entrez API. It is intended for literature discovery and retrieval in medicine, biology, and related life sciences. Search requests can combine ordinary keywords with MeSH terms, author names, date restrictions, and Boolean operators.

The server exposes three main retrieval capabilities. Search PubMed returns matching records, abstract access provides article abstracts and metadata through resource URIs, and full-text retrieval downloads the article content when PubMed Central makes it available. Full text is returned as XML, preserving document structure that can be useful for an AI assistant.

How it works

An MCP-compatible client starts the server as a Python module and communicates with it over the configured local process. The server sends requests to PubMed using Entrez, with an email address identifying the client as required by NCBI. An API key is optional but raises the documented request allowance from 3 requests per second to 10 requests per second.

The included prompts support three search workflows. Systematic Review Search helps assemble searches using MeSH terms, synonyms, and date filters. PICO Search structures a clinical question around population, intervention, comparison, and outcome. Author Search helps format an author query for finding that person’s publications. These prompts guide query construction; they do not expand the underlying PubMed database.

Setup and configuration

Install the package with pip install mcp-simple-pubmed. The server requires PUBMED_EMAIL; set PUBMED_API_KEY if higher NCBI rate limits are needed. A Claude Desktop configuration runs python -m mcp_simple_pubmed and passes these values in the server’s environment.

On macOS, Python installations may need their certificate bundle installed if HTTPS requests fail with a certificate verification error. The README identifies Python’s Install Certificates.command as the relevant fix and notes that Claude Desktop must be restarted after changing certificates.

Tools and capabilities

andybrandt/mcp-simple-pubmed MCP server supports:

  • PubMed searches using keywords, MeSH terms, authors, dates, and Boolean logic.
  • Abstract and metadata retrieval through MCP resource URIs.
  • Full-text retrieval for articles available through PubMed Central.
  • Prompt-assisted construction of systematic review, PICO, and author searches.

Limitations and notes

Full-text retrieval is limited by availability and access. The server cannot provide paywalled papers. A failed retrieval does not always prove that a paper is unavailable: browser access to a publication URL may work even when automated retrieval encounters a forbidden response.

The full-text response is XML rather than human-oriented prose. That format contains structural information, but applications should account for it when displaying or processing article content. No rate limiting is implemented by the server, so clients generating requests should stay within NCBI’s documented limits or use an API key where appropriate.

andybrandt/mcp-simple-pubmed MCP server is therefore best suited to PubMed search and accessible article retrieval, not unrestricted full-text access across all biomedical publishers.

Reviews

No reviews yet — be the first to share how it worked for you.

Frequently Asked Questions about MCP Simple Pubmed

andybrandt/mcp-simple-pubmed is an MCP server that connects AI assistants to PubMed through the NCBI Entrez API. Its main tools search articles, retrieve abstracts and metadata, and download available PubMed Central full text. It also provides prompts for systematic review, PICO, and author-focused search construction.

Install it with `pip install mcp-simple-pubmed`. Run it as the `mcp_simple_pubmed` Python module from an MCP client, and configure `PUBMED_EMAIL` in the client environment. Smithery installation is also documented for Claude Desktop.

The server requires `PUBMED_EMAIL`, while `PUBMED_API_KEY` is optional. NCBI requires the email address, and the API key provides the higher documented request limit of 10 requests per second instead of 3.

The README provides a standard Python MCP configuration for Claude Desktop, so use of the server with Cursor would require compatible MCP configuration that is not described here.

No, it cannot provide paywalled papers. Full-text retrieval works when content is available through PubMed Central, and an automated request may still fail even when the publication page opens in a regular browser.

Related MCP servers

Add a badge to your READMEShow it’s listed — live SVG, Markdown or HTML.

AllMCPs Directory Badge

Full Badge Customizer

Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.

Badge Style:
Live Dynamic SVG PreviewMCP Simple Pubmed AllMCPs Directory Badge
Markdown (GitHub README)
[![AllMCPs](https://allmcps.com/api/badge/andybrandt-mcp-simple-pubmed?style=directory)](https://allmcps.com/mcp/andybrandt-mcp-simple-pubmed)
HTML Embed
<a href="https://allmcps.com/mcp/andybrandt-mcp-simple-pubmed"><img src="https://allmcps.com/api/badge/andybrandt-mcp-simple-pubmed?style=directory" alt="MCP Simple Pubmed on AllMCPs" /></a>