# PubMed MCP > PubMed MCP gives your AI client direct access to the National Library of Medicine's database, pulling 37 million+ biomedical articles. Use it to search research by keyword or disease, instantly grab full article details from a PMID, and trace which papers cite specific research. It’s deep scientific literature retrieval, built for researchers and clinical writers. ## Overview - **Category:** the-unthinkable - **Price:** Free - **Tags:** biomedical-research, literature-search, citation-tracking, medical-database ## Description This MCP connects your AI agent directly to PubMed, the gold standard source for biomedical and life sciences literature. You can ask your client to find primary research using complex search terms like boolean operators (AND/OR/NOT) across tens of millions of articles. When you find a key paper, you don't have to navigate separate web pages; you simply request the full details, getting abstracts, all authors, and specific MeSH descriptors right away. Better yet, if that article is foundational, your agent can track down every subsequent study that cited it, mapping out how research evolves over time. It’s a massive shortcut for anyone who needs reliable scientific evidence quickly, giving you access to the entire catalog of tools hosted on Vinkius without needing an API key or manual data extraction. ## Tools ### get_pubmed_article Retrieves all detailed metadata for a specific article when you provide its unique PMID number. ### get_pubmed_citations Finds and lists articles that reference or cite a given PubMed paper, mapping the research impact chain. ### search_pubmed Search PubMed for biomedical articles, returning titles, abstracts, DOIs, and MeSH terms using AND, OR, or NOT operators. ## Prompt Examples **Prompt:** ``` Find recent research on CRISPR gene therapy for sickle cell disease. ``` **Response:** ``` Found 10 recent articles on CRISPR and sickle cell disease. Highlights include: 'CRISPR-Cas9 Gene Editing for Sickle Cell Disease and β-Thalassemia' (NEJM, 2023) — a landmark clinical trial showing 97% of patients free of vaso-occlusive crises after exa-cel infusion. PMID: 38040890. Full abstract and MeSH terms retrieved. ``` **Prompt:** ``` Get complete details for PubMed article PMID 33782455. ``` **Response:** ``` Retrieved full data for PMID 33782455: 'SARS-CoV-2 variants of concern and variants of interest' by Harvey et al. Published in Nature Reviews Microbiology (2021). The abstract details the classification framework for emerging variants including Alpha, Beta, Gamma, and Delta. DOI: 10.1038/s41579-021-00573-0. MeSH Terms: COVID-19, SARS-CoV-2, Mutation, Spike Glycoprotein. ``` **Prompt:** ``` Which papers cite the original CRISPR-Cas9 paper? Show me the top citing articles. ``` **Response:** ``` Found 10 citing articles for the landmark CRISPR paper. Top citations include applications in gene therapy, agricultural biotechnology, diagnostic systems (SHERLOCK/DETECTR), and base editing. Authors span institutions from MIT, Harvard, UC Berkeley, and the Broad Institute. Each result includes PMID, DOI, and full abstract. ``` ## Capabilities ### Search literature by topic Your agent executes targeted searches across millions of articles using keywords, disease names, and complex boolean logic. ### Retrieve full article metadata You get a comprehensive data dump for any specific paper, including the abstract, journal name, DOI, and author list. ### Track research impact The MCP identifies all later papers that built on or referenced a foundational piece of science. ## Use Cases ### Determining research gaps on a rare disease A researcher asks their agent to search_pubmed for all articles citing 'Gene X' combined with 'Disease Y'. The MCP returns the top 20 papers, allowing the researcher to instantly see which aspects of the disease are over-researched and where the gaps lie. ### Building a literature review on vaccine efficacy A writer needs evidence for a claim. They run search_pubmed using 'Vaccine A AND Efficacy' to find initial papers, then use get_pubmed_article with the PMID to pull the full abstract and authors for their bibliography. ### Tracking drug development over time A pharmacologist inputs a foundational paper on Drug Z. The agent uses get_pubmed_citations, showing all subsequent research that has built upon Drug Z's initial findings, demonstrating the evolution of treatment. ### Analyzing competing theories in oncology A clinician needs to compare two treatments. They use search_pubmed with complex boolean logic ('Treatment A OR Treatment B') AND 'Outcome Measure C', getting a highly focused list that eliminates irrelevant background noise. ## Benefits - You cut out hours of database clicking. Instead of navigating PubMed's complex web interface, you just tell your agent what you need, and it executes the search instantly. - When you find a key paper, you get full metadata immediately using get_pubmed_article. This means abstracts, every author, the DOI, and MeSH terms all in one clean output. - You can map out research history with ease. Use get_pubmed_citations to see which papers built on a landmark study—critical for academic reviews. - Complex search queries are simple. The search_pubmed tool handles boolean operators (AND, OR, NOT), letting you narrow results using precision language instead of keyword stuffing. - It works with your existing workflow. Whether it's Claude or Cursor, your agent sends the scientific query and gets structured data back for analysis. ## How It Works The bottom line is you get reliable, complex scientific literature analysis without leaving your chat window or IDE. 1. Connect your AI client to this MCP in the Vinkius catalog. No API key is needed. 2. Ask your agent to perform a search, providing keywords or a PMID. Your agent sends the query directly to PubMed. 3. The system returns structured data containing abstracts, author details, and citation lists for you to use. ## Frequently Asked Questions **How do I use the PubMed MCP to find papers on a specific disease?** Use search_pubmed and include keywords like 'Disease Name AND Gene Symbol'. The tool supports powerful boolean operators, so you can narrow down results precisely. **Can I get all author names from PubMed using the MCP?** Yes. After finding a relevant PMID, calling get_pubmed_article retrieves comprehensive metadata that includes every contributing author and their affiliation. **What is the difference between search_pubmed and get_pubmed_article?** search_pubmed finds many articles based on keywords. get_pubmed_article pulls all the granular details for one specific article if you already have its PMID. **How do I check which papers cite a landmark study?** Use get_pubmed_citations, passing in the foundational paper's PMID. This tool maps out the entire impact chain and shows follow-up research. **Do I need to provide an API key for PubMed MCP?** No. Since this is hosted on Vinkius, you connect your preferred AI client once, and no keys or complex setup are required to access the database.