# pubcrawl [Health: Active]

**Category:** 🧬 Biology & Bioinformatics  
**Repository:** https://github.com/nickjlamb/pubcrawl  
**GitHub Stars:** 15  
**npm Downloads (last month):** 1350  
**Views:** 4  
**Installs:** 0  
**Upvotes:** 0  
**Directory Page:** https://allmcps.com/mcp/pubcrawl

## Description
PubMed literature search and US/UK drug labelling for AI assistants.

## Claude Desktop Quick Installation
Install path detected from listing signals. Uses `npx` (confidence: high):

```json
"mcpServers": {
  "pubcrawl": {
    "command": "npx",
    "args": ["-y","@pharmatools/pubcrawl"]
  }
}
```

## Documentation

## What pubcrawl MCP server does

The pubcrawl MCP server connects an MCP-compatible AI assistant to biomedical information sources used by researchers and clinicians. Its tools cover three areas: literature, drug labelling, and clinical trials. Results include source identifiers such as PMIDs, NCT IDs, and DOIs where available, allowing retrieved information to be checked against the originating service.

Literature tools search PubMed and Europe PMC, retrieve structured abstracts or open-access full text, find related articles, format citations, and identify recent or high-impact papers. Europe PMC searches can include preprints and patents, with filters for preprints and open-access content. Drug tools resolve generic and brand names, retrieve US Prescribing Information and UK Summaries of Product Characteristics, compare those labels, and search for drugs approved for a condition. Trial tools search ClinicalTrials.gov and retrieve detailed records by NCT ID.

This makes the pubcrawl MCP server suitable for evidence discovery, regulatory copy review, literature preparation, and trial screening in the Biology, Medicine and Bioinformatics category.

## How it works

Each tool is described as a thin wrapper around an official external API. The MCP client sends a natural-language request to the selected tool, and the server returns data from the relevant source rather than generating a substitute record. The available sources include PubMed, Europe PMC, openFDA, DailyMed references, the UK electronic Medicines Compendium, and ClinicalTrials.gov.

The server can also support cross-source workflows. For example, an assistant can retrieve a drug label, compare US and UK sections, and then search for related clinical trials or publications. Label comparison maps corresponding sections so differences in indications, warnings, contraindications, or dosing can be reviewed side by side.

## Setup and configuration

Add the package to an MCP client configuration using `npx`; the client fetches it on first use, so a separate installation step is not required. A Claude Desktop configuration uses the `@pharmatools/pubcrawl` package with the `pubcrawl` server name and the `npx -y` arguments. After saving the configuration, restart the client to connect the server.

No API key is required for the standard setup.

## Tools and capabilities

- Search PubMed with date, article-type, and sort filters.
- Search Europe PMC for articles, preprints, and patents.
- Retrieve abstracts, open-access full text, related papers, and formatted citations.
- Resolve drug names and retrieve or compare US and UK labels.
- Search ClinicalTrials.gov and inspect trial eligibility, design, outcomes, locations, and related publications.

## Limitations and notes

Full-text retrieval is described for open-access articles from PubMed Central; the material does not indicate access to paywalled full text. Drug-label and trial results depend on the underlying external sources. The optional NCBI key only affects PubMed rate limits, and no key is needed for the basic configuration.

The pubcrawl MCP server is reported as compatible with Claude Desktop and Cursor, as well as other MCP-compatible clients. The provided setup instructions specifically document Claude Desktop configuration. The project is licensed under MIT.

_Full upstream README: https://allmcps.com/mcp/pubcrawl/readme_