RAG Web UI
A retrieval-augmented-generation dialogue system for building a knowledge-base question-answering service from your own documents, with a vector-database-backed retrieval engine.
RAG Web UI is an intelligent dialogue system based on retrieval-augmented generation that helps build a knowledge-base question-answering service from a user's own documents. It combines document retrieval with a large language model to produce cited answers.
It supports multiple document formats, including PDF, DOCX, Markdown and plain text, with automatic chunking and vectorization and asynchronous, incremental processing of new documents. The dialogue engine supports multi-turn conversation with reference citations in its answers. It can use cloud language models such as OpenAI, DeepSeek and MiniMax, or local models through Ollama, and exposes OpenAPI interfaces for programmatic access to the knowledge base. The architecture separates frontend and backend, uses distributed file storage, and lets a vector database be switched between ChromaDB and Qdrant through a factory pattern.
It deploys with Docker Compose, requiring Node.js 18+, Python 3.9+ and at least 8GB of RAM, and is written in TypeScript under the Apache 2.0 license.