Integrations
Provides containerized deployment of the PubTator MCP server with a ready-to-use Dockerfile for easy setup and isolation
Requires Python 3.10+ to run the server, with explicit configuration examples and installation instructions for Python environments
Supports exporting PubTator annotation results in biocxml format for biomedical literature annotations
PubTator MCP Server
🔍 A biomedical literature annotation and relationship mining server based on PubTator3, providing convenient access through the MCP interface.
PubTator MCP Server provides AI assistants with access to the PubTator3 biomedical literature annotation system through the Model Context Protocol (MCP). It allows AI models to programmatically search scientific literature, obtain annotation information, and analyze entity relationships.
🤝 Contribute • 📝 Report Issues
✨ Core Features
- 🔎 Literature Annotation Export: Support exporting PubTator annotation results in multiple formats ✅
- 🚀 Entity ID Lookup: Query standard identifiers for biological concepts through free text ✅
- 📊 Relationship Mining: Discover biomedical relationships between entities ✅
- 📄 Literature Search: Support literature retrieval by keywords and entity IDs ✅
- 🧠 Batch Processing: Support batch export of annotation information from search results ✅
🚀 Quick Start
Requirements
- Python 3.10+
- FastMCP library
Installation
Via Smithery
Use Smithery to automatically install PubTator Server:
Claude
Cursor
Paste in Settings → Cursor Settings → MCP → Add new server:
- Mac/Linux
Windsurf
CLine
Manual Installation
- Clone the repository:Copy
- Install dependencies:Copy
📊 Usage
Running the Server Directly
Start the MCP server:
The server now supports both stdio and TCP transports. By default, it uses TCP transport. You can configure the following environment variables:
MCP_TRANSPORT
: Set to "tcp" for TCP transport or "stdio" for stdio transport (default is "tcp")MCP_HOST
: The host to bind to (default is "0.0.0.0")MCP_PORT
: The port to listen on (default is 8080)
Example of starting the server with custom settings:
The server implements lazy initialization and proper error handling. It will gracefully handle shutdown signals (SIGINT and SIGTERM) and log any errors that occur during startup or operation.
Using Docker
We provide a Dockerfile for easy deployment. To use the Docker container:
- Build the Docker image:Copy
- Run the Docker container:Copy
This will start the PubTator MCP server inside a Docker container, exposing it on port 8080.
Troubleshooting
If you encounter any issues starting the server:
- Check the console output for error messages.
- Ensure all required dependencies are installed (see Requirements section).
- Verify that the environment variables are set correctly.
- If the server fails to start, try running it with increased verbosity:
This will provide more detailed logging information to help identify the source of any issues.
When using Docker, you can check the logs with:
Configuration
Claude Desktop Configuration
Add to claude_desktop_config.json
:
(Mac OS)
(Windows)
CLine Configuration
To use TCP transport, modify the configuration as follows:
🛠 API Features
PubTator MCP Server provides the following core features:
1. Export Publications (export_publications)
Export PubTator annotation results for specified PMID literature:
2. Entity ID Lookup (find_entity_id)
Query standard identifiers for biological concepts through free text:
3. Relationship Query (find_related_entities)
Find other entities related to a specified entity:
4. Literature Search (search_pubtator)
Search the PubTator database:
5. Batch Export (batch_export_from_search)
Search and batch export literature annotations:
Note: The actual function calls may vary depending on your implementation. These examples are based on our recent tests and may need to be adjusted to match your exact API.
⚠️ Usage Limitations
- API request rate limit: maximum 3 requests per second
- When batch exporting, use a reasonable batch_size to avoid request timeout
- For relationship queries, entity IDs must start with "@", e.g., "@DISEASE_COVID-19"
📄 License
This project is licensed under the MIT License.
⚠️ Disclaimer
This tool is for research purposes only. Please comply with PubTator's terms of service and use this tool responsibly.
This server cannot be installed
remote-capable server
The server can be hosted and run remotely because it primarily relies on remote services or has no dependency on the local environment.
🔍 A biomedical literature annotation and relationship mining server based on PubTator3, providing convenient access through the MCP interface.
Related MCP Servers
- AsecurityAlicenseAqualityAn MCP server that provides access to PubMed articles through the Entrez API.Last updated -228PythonMIT License
- -securityAlicense-qualityServer to search PubMed (PubMed is a free, online database that allows users to search for biomedical and life sciences literature). I have created on a day MCP came out but was on vacation, I saw someone post similar server in your DB, but figured to post mine.Last updated -11PythonMIT License
- -securityFlicense-qualityAn MCP server implementation that enables searching and retrieving research articles from PubMed with specific focus on open access content filtering and full-text link retrieval.Last updated -43JavaScript
- -securityFlicense-qualityEnables search and retrieval of academic papers from PubMed database with advanced features like MeSH term lookup, publication statistics, and PICO-based evidence search.Last updated -Python