PubCrawl

You can run PubCrawl, an MCP server that grants LLM clients access to PubMed, drug labeling data for the US and UK, and ClinicalTrials information. It is designed to be connected by MCP clients so you can query literature, regulatory labeling, and trials in a single, cohesive workflow.

How to use

Connect PubCrawl to an MCP client, then issue natural language prompts to access literature, drug labeling, and clinical trials data. You can search PubMed with filters, fetch abstracts or full texts, compare US and UK labeling, retrieve UK SmPCs, and explore trials by condition or phase. Use natural language questions like “What are the latest trials on semaglutide?” or “Get the US prescribing information for metformin, indicators and warnings.” You can also combine tools to format citations, find related articles, or pull recent papers on a topic. When you are connected, PubCrawl serves as a centralized data source for medical literature, labeling information, and clinical trials through your MCP client interactions.

How to install

Prerequisites you need before installing PubCrawl include Node.js 20+ and an MCP-compatible client such as Claude Desktop or Cursor.

Install PubCrawl from npm globally.

npm install -g @pharmatools/pubcrawl

Additional setup details

Configure Claude Desktop to connect PubCrawl by adding an MCP server entry that points to the local command. Start PubCrawl after building from source as described below.

Configuration and runtime notes

If you are building from source, PubCrawl runs as a local process. After building, you start the server with the Node runtime and point to the built index file.

Troubleshooting and tips

If you encounter rate limits from external data sources, you can supply an API key when supported to increase request quotas. Always ensure your Node.js version matches the prerequisites and that your MCP client is configured to connect to PubCrawl correctly.

Notes on usage patterns

PubCrawl exposes tools for literature searches, abstract retrieval, full-text access from PubMed Central, related article discovery, and formatted citations. It also provides access to US and UK drug labeling information and ClinicalTrials details, enabling comprehensive research workflows within a single MCP session.

Available tools

search_pubmed

Search PubMed with filters for date range, article type, and sort order. Returns PMIDs, titles, authors, journals, and DOIs.

get_abstract

Retrieve the full structured abstract for an article, including background, methods, results, conclusions, plus keywords and MeSH terms.

get_full_text

Fetch the full text of open-access PubMed Central articles, with parsed sections, figure/table captions, and reference counts.

find_related

Find similar articles using PubMed's neighbor algorithm, ranked by relevance score.

format_citation

Generate a formatted citation in APA, Vancouver, Harvard, or BibTeX style.

trending_papers

Find recent papers on a topic, with optional filtering to high-impact journals.

get_uspi

Pull US Prescribing Information sections from DailyMed — indications, dosing, warnings, contraindications, and more.

get_smpc

Retrieve UK Summary of Product Characteristics from the eMC — the UK labeling with numbered sections.

compare_labels

Side-by-side comparison of USPI and SmPC labeling for the same drug to spot regulatory differences.

search_by_indication

Find drugs approved for a medical condition by querying FDA labeling and UK availability.

search_trials

Search ClinicalTrials.gov for trials with filters like condition, intervention, status, and phase.

get_trial

Get full trial details by NCT ID, including eligibility, design, outcomes, and locations.