{"repo":"knows-cloud/paperless-iq","free":true,"listed":false,"github":"https://github.com/knows-cloud/paperless-iq","clone":"git clone https://github.com/knows-cloud/paperless-iq.git","description":"Open-source AI add-on for Paperless-NGX: automatic metadata & tagging, RAG document search, and automation with Ollama, Bedrock, Anthropic, and OpenAI","language":"Python","stars":15,"topics":["ai","amazon-bedrock","anthropic","automation","bedrock","docker","document-analysis","document-classification","document-management","document-search"],"license":null,"category":"workflow-automation","readme_excerpt":"Paperless IQ Add LLM intelligence to your Paperless-NGX — self-hosted, privacy-first, and built for non-English archives. Paperless IQ connects to your existing Paperless-NGX instance and gives it a brain: an LLM reads every document, suggests structured metadata, and lets you have natural conversations with your entire archive. No cloud dependency, no data leaving your server. --- How Paperless IQ compares to Paperless-NGX v3 Paperless-NGX v3.0 (currently in beta) ships its own LLM integration: metadata suggestions for titles, tags, correspondents, and document types, plus a document chat feature backed by FAISS vector search. That's a meaningful step forward from the old rule-based classifier. Paperless IQ goes deeper on every dimension — more providers, more control, stronger retrieval, and features that Paperless-NGX v3 simply doesn't have. Paperless-NGX v3.0 (beta) Paperless IQ --- --- --- LLM metadata suggestions Titles, dates, tags, correspondents, doc types All of those + custom fields LLM providers OpenAI-compatible APIs · Ollama Ollama · Anthropic · OpenAI · Amazon Bedrock (Claude, Nova, Llama, Mistral) Approval workflow Inline suggestions Full queue — editable fields, suggestion stacking, batch approve/reject Scanned / image-only PDFs Azure AI cloud OCR On-premise vision analysis — LLM renders and reads pages directly Document chat / RAG Yes, via FAISS (\"hit-or-miss\" per community) ChromaDB · Qdrant with hybrid dense+sparse search · Bedrock KB Long-term memory — Fa","default_branch":null,"files":null,"tree":[],"storefront":"/r/knows-cloud","claimed":false,"request_supported":{"post":"https://gitbuyer.com/r/knows-cloud/paperless-iq/request-supported","requests":0},"note":"indexed from public GitHub; nothing is for sale on this page. Clone it from GitHub. Paid listings live at /search."}