Turn your Paperless-ngx into an AI-powered document assistant that auto-tags PDFs and lets you chat with your entire archive using RAG - supports OpenAI, Ollama, and 10+ other AI backends

Ever wished you could just ask your document archive “What was my last electricity bill amount?” instead of hunting through folders? Paperless-AI bridges that gap by supercharging Paperless-ngx with AI-powered document understanding. It automatically analyzes new documents, assigns smart tags and metadata, then builds a searchable knowledge base you can chat with in natural language.

What sets this apart is its flexibility - it works with OpenAI, but also Ollama for local models (Mistral, Llama, Phi-3), plus DeepSeek, Perplexity, and basically any OpenAI-compatible API. The RAG implementation means you’re not just doing keyword searches, but actually understanding document context. With 5140+ stars, it’s clearly solving a real problem for the self-hosted community.

Perfect for anyone running Paperless-ngx who wants to level up their document workflow without vendor lock-in. Docker deployment is straightforward, and the web interface handles both automated processing and manual review for sensitive docs.

⭐ Stars: 5140
💻 Language: JavaScript
🔗 Repository: clusterzx/paperless-ai