pubmed-search

PubMed Search

Search and analyze biomedical literature from PubMed using the free NCBI E-utilities API.

When to Use

• Discovering biomedical or clinical papers by keyword, author, journal, or date range
• Fetching structured metadata (title, authors, abstract, DOI) for a known PMID
• Performing deep analysis of a paper when only the abstract and metadata are available
• Downloading open-access full-text PDFs from PubMed Central (PMC)
• Building a candidate list before running triage-paper

When Not to Use

• The paper is already known (DOI, URL) — go straight to triage-paper
• A semantic-scholar MCP or PubTator MCP is configured — prefer the MCP; it returns structured data with no rate-limit risk
• The search is for general academic literature — use google-scholar-search or semantic-scholar-search
• A candidate JSON already exists at /tmp/<topic>-candidates.json — reuse it

Recommended MCP Server

When available, prefer the PubTator MCP server over this script:

{
  "mcpServers": {
    "pubtator": {
      "type": "stdio",
      "command": "uvx",
      "args": ["pubtator-mcp-server"]
    }
  }
}

Mindset

Search is discovery, not analysis. The goal is a structured candidate list.

1. Rate limits are a gotcha: without an API key the limit is 3 req/s; a pitfall is issuing bulk PMID fetches without a delay, causing silent failures or 429s. ALWAYS add a short delay between batch calls.
2. MCP first: ALWAYS check whether a PubTator or PubMed MCP is configured before invoking the Python script. MCPs are faster, structured, and avoid rate-limit risk.
3. Open access is not guaranteed: a pitfall is assuming all PMC articles can be downloaded. NEVER attempt to download a PDF without first confirming PMC availability and open-access status.

Workflow

1. Check MCP availability

ALWAYS check for a pubtator or pubmed MCP before running the script. If configured and reachable, prefer it.

2. Set up the environment (first run only)