PubMed & ClinicalTrials.gov MCP Server

MCP (Model Context Protocol) server providing AI agents with direct access to 37M+ PubMed articles and 500K+ clinical trials from ClinicalTrials.gov — two official NIH data sources in one server.

Connect any MCP-compatible client (Claude Desktop, Cursor, Cline, etc.) to search medical research and clinical studies.

Tools

Tool	Description
`search_pubmed`	Search PubMed articles by keyword, with full abstracts
`get_article_abstracts`	Fetch abstracts for specific PMIDs
`search_clinical_trials`	Search ClinicalTrials.gov studies
`get_article_details`	Get detailed metadata for a PubMed article

Connect via MCP

Claude Desktop / Cursor

{
  "mcpServers": {
    "pubmed": {
      "url": "https://aligned-kettledrum--pubmed-mcp-server.apify.actor/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_APIFY_TOKEN"
      }
    }
  }
}

Using mcp-remote (npx)

npx mcp-remote https://aligned-kettledrum--pubmed-mcp-server.apify.actor/mcp \
  --header "Authorization: Bearer YOUR_APIFY_TOKEN"

Example Queries

Once connected, ask your AI assistant:

"Find recent papers about CRISPR gene editing in cancer therapy"
"Search for clinical trials on GLP-1 receptor agonists for obesity"
"Get the abstract for PMID 37198476"
"What are the latest Phase 3 trials for Alzheimer's disease?"

Data Source

PubMed: NCBI E-utilities API — 37M+ biomedical articles from 5,200+ journals
ClinicalTrials.gov: v2 API — 500K+ clinical studies worldwide

Both APIs are maintained by the U.S. National Institutes of Health. Free, no API key required.

FAQ

Q: Do I need an API key?A: No. Both PubMed and ClinicalTrials.gov APIs are free and open.

Q: How current is the data?A: PubMed indexes new articles within days of publication. ClinicalTrials.gov updates are near real-time as sponsors submit data.

Q: Can I get full-text articles?A: This server provides abstracts and metadata. Full-text access depends on publisher open-access policies.

Q: Is there a rate limit?A: The server handles NCBI's rate limiting (3 req/sec) automatically. No client-side throttling needed.

Batch Processing

For large-scale data extraction (thousands of articles), use the PubMed & ClinicalTrials.gov Actor which supports bulk downloads with pagination.

PubMed & ClinicalTrials.gov MCP Server