Welcome to the BMLibrarian documentation! BMLibrarian is a comprehensive Python library and application providing AI-powered access to biomedical literature databases, featuring a sophisticated multi-agent architecture for research automation.

• Getting Started - Installation and first steps
• User Guide - Complete guide to using BMLibrarian
• CLI Reference - Command-line interface documentation
• Qt GUI Guide - Desktop application user guide

• Architecture Overview - System design and components
• Plugin Development Guide - Create custom plugins for the Qt GUI
• API Reference - Python API documentation
• Contributing Guidelines - How to contribute to the project
• Development Setup - Set up your development environment

• Research Workflow - Multi-agent research automation
• Fact Checker Guide - Validate biomedical statements
• Query Optimization - Get better search results

BMLibrarian is a comprehensive Python library that provides AI-powered access to biomedical literature databases. It features a multi-agent architecture with specialized agents for:

• Query Processing - Convert natural language questions to database queries
• Document Scoring - Assess relevance of documents to research questions
• Citation Extraction - Extract specific passages that answer questions
• Report Generation - Synthesize evidence into publication-quality reports
• Counterfactual Analysis - Find contradictory evidence for balanced analysis
• Fact Checking - Validate biomedical statements against literature

• Specialized AI agents for different research tasks
• Coordinated workflow orchestration
• Queue-based batch processing
• Support for multiple LLM models (via Ollama)

• PostgreSQL with pgvector for semantic search
• Support for PubMed (38M+ articles) and medRxiv preprints
• Full-text search and vector similarity search
• Automatic schema migrations

• CLI - Interactive command-line interface
• Qt GUI - Modern desktop application with plugin architecture
• Python API - Full programmatic access

• Multi-model query generation (20-40% more relevant documents)
• Citation hallucination prevention
• Query performance tracking
• OpenAthens institutional access support
• PDF management and viewing

• Runs entirely with local models (Ollama)
• No API keys required
• No proprietary services needed
• Complete offline operation (after initial data sync)

• Python: 3.12 or higher
• Database: PostgreSQL 14+ with pgvector extension
• AI/LLM: Ollama for local model inference
• Package Manager: uv (recommended) or pip

# Install dependencies
uv sync

# Set up database and import data
uv run python initial_setup_and_download.py test_database.env

# Run the Qt GUI
uv run python bmlibrarian_qt.py

# Or run the CLI
uv run python bmlibrarian_cli.py

See the Getting Started guide for detailed installation instructions.

bmlibrarian/
├── src/bmlibrarian/          # Main source code
│   ├── agents/               # Multi-agent system
│   ├── importers/            # Data importers (PubMed, medRxiv)
│   ├── embeddings/           # Document embeddings
│   ├── factchecker/          # Fact-checking system
│   ├── gui/qt/               # Qt GUI application
│   │   ├── plugins/          # Plugin-based architecture
│   │   ├── widgets/          # Reusable UI components
│   │   └── resources/        # Styles and assets
│   ├── database.py           # Database access layer
│   └── config.py             # Configuration management
├── doc/                      # Comprehensive documentation
│   ├── users/                # End-user guides
│   └── developers/           # Technical documentation
├── tests/                    # Test suite
└── examples/                 # Example scripts

• Getting Started - Installation and setup
• User Guide - Using BMLibrarian for research
• CLI Reference - Command-line tools
• Qt GUI Guide - Desktop application
• Fact Checker Guide - Statement validation
• Configuration Guide - Customize BMLibrarian

• Architecture Overview - System design
• Plugin Development Guide - Create Qt plugins
• Agent Development - Create custom agents
• Database Schema - PostgreSQL schema reference
• API Reference - Python API documentation
• Testing Guide - Writing and running tests

• Research Workflow - Automated research process
• Query Optimization - Improve search results
• Citation Best Practices - Extract quality evidence
• Multi-Model Queries - Use multiple AI models

• GitHub Repository: hherb/bmlibrarian
• Issue Tracker: Report bugs and request features
• Discussions: Ask questions and share ideas
• Contributing: See Contributing Guidelines

November 2025:

• ✅ Qt GUI with plugin architecture
• ✅ Fact Checker system with CLI and GUI
• ✅ Multi-model query generation
• ✅ Query performance tracking
• ✅ Citation hallucination prevention
• ✅ PostgreSQL audit trail
• ✅ Automatic database migrations

BMLibrarian is open-source software. See the repository for license details.

Next Steps:

1. Install BMLibrarian and set up your environment
2. Follow the User Guide to learn the basics
3. Try the Qt GUI for a visual research workflow
4. Develop a Plugin to extend functionality

Welcome to BMLibrarian! 🎉