PubMed MCP Server by cyanheads
PubMed MCP searches biomedical papers, retrieves available full text and formats citations with retraction and correction notices.
- 167 ★ on the repository
- Last commit
- Apache-2.0
Extended access
Why this level
- Search topics and identifiers reach literature services and, for hosted use, the server operator.
Install
- Works in
Manual install
claude mcp add --transport http pubmed-mcp-server https://pubmed.caseyjhand.com/mcpFirst Getting started option, hosted HTTP without local installation. The next README block covers local execution.
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Search topics and identifiers reach literature services and, for hosted use, the server operator.
This is third-party code. Review the repository files before installing.
What it does
PubMed MCP Server helps an agent search PubMed and Europe PMC and retrieve abstracts, authors and publication metadata. Full text comes from open sources, while linked correction and retraction notices help assess findings. Tools find related work, convert DOI, PMID and PMCID identifiers, look up MeSH subject headings and format citations. Connect through the public endpoint or a local process. Assessing study quality and interpreting evidence remain the reader’s and agent’s responsibility.
Who it is for. Biomedical researchers building literature sets and checking publication links and notices through an agent client.
Good fit when
- Build a literature search with date and publication-type filters.
- Fetch an abstract and linked notices by PMID.
- Format a mixed set of references into a bibliography.
- Refine a query with MeSH and related publications.
Not a fit when
- Every paywalled article must be available in full.
- A single search result would determine clinical action without expert assessment.
- You need comprehensive coverage of all disciplines beyond biomedicine.
Example request
Collect publications for a topic and date range, save PMIDs and links, and mark retractions and unavailable full text.MCP
- Transport
- http, stdio
- Authentication
- not required
Limitations
Full text is available only where connected sources provide it; the server reports unavailable articles and truncated responses. NCBI rate limits apply, and an optional key raises permitted throughput without removing limits. The Unpaywall full-text fallback needs a separately configured contact email. Public-endpoint queries pass through third-party infrastructure; for sensitive topics, consider local execution and the rules governing disclosure of search queries.
How to disable. Remove the PubMed MCP entry and stop any local process. Revoke an unneeded NCBI key separately.
FAQ
Can I connect without installation?
Yes. Use the documented public HTTP server and consider where queries are processed.
Is an NCBI key required?
Not to start. Configure `NCBI_API_KEY` for higher throughput while respecting service limits.
Where does full text come from?
PMC, then Europe PMC and configured Unpaywall. Check the returned source and license.
Are retractions indicated?
Yes. Article data and formatted citations include linked notices; inspect them before using the study.
What is MeSH?
A medical subject vocabulary for refining searches. Look up a heading and inspect its definition.
Discussion
No questions or reviews yet.
Related
MCP serversMedical MCP
medical-mcp
MCP server for medical sources: FDA and EMA drug data, WHO statistics, PubMed and clinical trials without API keys.
An MCP server for searching and downloading academic papers from arXiv, PubMed, bioRxiv, Semantic Scholar and other open sources
BioMCP
biomcp
One binary and a single command grammar for 70 biomedical sources: PubMed, ClinVar, ClinicalTrials.gov, OncoKB. MCP mode starts with the serve command
Connects a Zotero library to Claude and other agents: search papers, annotate, cite, and search by meaning
An MCP server and skill for Sciverse paper search: metadata, semantic retrieval, citations and full text for agents.
An MCP server gives agents search and answers over local documents using hybrid retrieval, reranking and section links.
README in short
Responses expose effective queries, provenance and partial-failure reasons. A research-plan prompt structures the work into phases.