zenpdf

Local PDF Q&A with RAG using Ollama & LangChain

A peaceful CLI tool for chatting with your documents using local AI models. All processing happens on your machine - no cloud APIs, no data leaves your device.

Features

• 🔒 Local-First - No cloud APIs, all processing on your machine
• 📄 Multi-Format - PDF, DOCX, and TXT support
• ⚡ Streaming - Real-time AI responses as they're generated
• 💾 Persistent History - Chat history saved between sessions
• 📚 Source Attribution - Know which documents informed each answer
• 🎨 Beautiful CLI - Rich terminal interface with colors and tables
• ⚙️ Fully Configurable - Customize models, chunk sizes, and more

Installation

pip install zenpdf

Quick Start

# 1. Make sure Ollama is running
ollama

# 2. Pull required models (if not already done)
ollama pull llama3.2:1b
ollama pull nomic-embed-text

# 3. Index a document
zenpdf index ./my-document.pdf

# 4. Ask questions!
zenpdf ask "What is this document about?"

# 5. Interactive mode
zenpdf interactive

Commands

Document Operations

Query Operations

Reference & History

Configuration

Utilities

Configuration

Default settings (view with zenpdf config show):

Configuration is saved to .zenpdf_config.json in your working directory.

Requirements

• Python 3.11+
• Ollama installed and running
• Ollama models:
  • llama3.2:1b (or your preferred model)
  • nomic-embed-text (for embeddings)

Architecture

┌─────────────┐     ┌──────────────┐     ┌─────────────┐
│  Document   │────▶│   Splitter   │────▶│   ChromaDB  │
│   Loader    │     │   (Chunks)   │     │   (Vectors) │
└─────────────┘     └──────────────┘     └──────┬──────┘
                                                 │
                      ┌──────────────┐           │
                      │   Ollama     │◀──────────┘
                      │  (Embeddings)│
                      └──────┬───────┘
                             │
                      ┌──────▼───────┐
                      │  RAG Chain   │
                      └──────┬───────┘
                             │
                      ┌──────▼───────┐
                      │      LLM     │
                      │   (Ollama)   │
                      └──────────────┘

Tech Stack

• LangChain - LLM orchestration
• Chroma - Vector database
• Ollama - Local LLMs
• Click - CLI framework
• Rich - Terminal formatting

License

MIT License - see LICENSE

Contributing

Contributions welcome! Please open an issue or submit a PR.

Made with ❤️ for local AI