kotaemon

Why we included this project

kotaemon is a web interface for asking questions over your own documents, and it skips the usual assembly work: the retrieval stack ships built in, with a hybrid retriever that blends full-text and vector search plus re-ranking. Answers come with citations you can check inside the built-in PDF viewer, which matters when you need to trust what the model says. It runs against hosted LLM APIs or local models via Ollama and llama-cpp, so teams that want on-premise document QA can keep everything in-house. Developers get more than a finished app: the codebase is a RAG pipeline framework with a Gradio UI to extend, and it handles multi-modal parsing and agent-style reasoning for multi-hop questions. For a small team that wants a working document assistant without building one, this is a solid, customizable starting point.