lemma-ai 2.0.0

Local-first paper manager with semantic search and LLM reasoning

📚 Lemma - Local-First Research Paper Manager

A privacy-first research paper manager with local semantic search and optional AI-powered insights.

✨ Features

• 🔒 Privacy-First: All papers stored locally, no cloud uploads
• 🚀 Fast Semantic Search: Local vector search across all papers
• 🤖 AI Q&A (Optional): Ask questions using cloud LLMs
• 📊 Auto-Processing: One command to scan, rename, and embed
• 🔄 Incremental Updates: 70-90% faster re-embedding
• 📂 Smart Cleanup: Automatically removes deleted papers from database
• 👀 Watch Mode: Auto-process new papers as they're added

🚀 Quick Start

1. Installation

pip install -r requirements.txt

2. Set Your Papers Folder (One-Time Setup)

# Set default papers directory and process all papers
lemma sync ~/Papers --set-default

# That's it! Your papers are now:
# ✓ Scanned and indexed
# ✓ Renamed with metadata
# ✓ Embedded for semantic search
# ✓ Ready for questions

3. Add New Papers (Automatic)

Option A: Manual Sync

# Just drop PDFs into ~/Papers, then run:
lemma sync

Option B: Auto-Sync (Watch Mode)

# Start watching (leave running in terminal)
lemma sync --watch

# Now just drop PDFs into ~/Papers
# They're automatically processed in seconds!

4. Query Your Papers

lemma ask "What are the main findings?"

📖 Common Workflows

First-Time Setup

# 1. Set your papers folder and sync everything
lemma sync ~/Papers --set-default

# 2. Query your papers
lemma ask "What is the main contribution?"

Daily Use

# Download new papers to ~/Papers, then:
lemma sync

# Or enable auto-processing:
lemma sync --watch  # Leave running

Browse Your Library

lemma list                    # List all papers
lemma search "transformers"   # Search by keyword
lemma show 5                  # Show paper details

🔧 Advanced Usage

Sync Options

lemma sync                    # Use default directory
lemma sync ~/Papers           # Specify directory
lemma sync --no-rename        # Skip automatic renaming
lemma sync --no-embed         # Skip embedding (faster)
lemma sync --watch            # Continuous monitoring

Manual Control (If Needed)

lemma scan ~/Papers           # Just scan (no rename/embed)
lemma organize                # Rename existing files
lemma embed                   # Generate embeddings only
lemma embed-status            # Check embedding coverage
lemma verify --remove         # Clean up missing files

🤖 AI Q&A Setup (Optional)

Set up an API key to enable question answering:

# Option 1: Environment variable
export GROQ_API_KEY="your_key_here"

# Option 2: .env file
echo "GROQ_API_KEY=your_key_here" > ~/.lemma/.env

# Then ask questions
lemma ask "What are the main approaches discussed?"

Get Free API Keys:

• Groq - Fast and generous free tier (recommended)
• Google Gemini - Alternative option

📋 Key Commands

Command	Description
lemma sync	Auto-process papers (scan + rename + embed)
lemma sync --watch	Monitor folder and auto-process new papers
lemma list	List all indexed papers
lemma ask <question>	Ask questions across papers (requires API key)
lemma search <query>	Search papers by keyword
lemma show <id>	Show paper details
lemma embed-status	Check embedding coverage

🔒 Privacy

• All papers stay on your machine - never uploaded anywhere
• Embeddings generated locally - no external API calls
• Cloud APIs only for Q&A - and only if you configure them
• Database stored locally at ~/.lemma/lemma.db

📦 Requirements

• Python 3.10 or higher
• ~500MB disk space for embedding models
• Internet connection only for optional AI Q&A

License

MIT License - see LICENSE file for details

Support

• Report issues: GitHub Issues
• Questions: Open a discussion on GitHub