RAG Web UI
A self-hosted document Q&A app with source citations, local models through Ollama and optional cloud models from OpenAI, DeepSeek and MiniMax.
RAG Web UI is a self-hosted app for asking questions about your own documents. It's a learning project, and its maintainers advise against commercial use and state that it isn't ready for production.
- Document knowledge bases: Import PDF, DOCX, Markdown and text files, with automatic preparation for search, background processing and incremental updates.
- Answers with references: Ask follow-up questions with conversation context and check citations against the retrieved source material.
- Local or cloud models: Use Ollama for local model generation, or send generation requests to OpenAI, DeepSeek or MiniMax cloud services.
- Knowledge base API: Access knowledge bases from other applications through an API, alongside the browser interface for managing documents and chatting with streaming responses.
The app supports Docker deployment and requires at least 8 GB of RAM. It supports ChromaDB or Qdrant for document search and MinIO for file storage. Choosing a cloud model sends generation requests to that provider even when you host the app yourself. The code is open source under Apache 2.0.
