KnowledgeBase-RAG-LLM-System
A learning project that implements a local knowledge base and retrieval-augmented question answering system with Streamlit, LangChain and Chroma, letting a user upload text files and then chat with answers grounded in that uploaded content.
KnowledgeBase-RAG-LLM-System is built as an entry-level, hands-on practice project for local knowledge base question answering and retrieval-augmented generation. One Streamlit page lets a user upload text files, which are split into chunks with a recursive character splitter, deduplicated by MD5 hash, and written into a locally persisted Chroma vector store; a second Streamlit page provides a chat interface that retrieves relevant chunks and passes them through a LangChain retrieval-prompt-LLM-output chain, with streaming output and message history stored to file.
It uses the DashScope text-embedding-v4 embedding model and the Qwen3-max chat model by default, configured through a Python settings file, and ships with sample content about clothing sizing and care as a demonstration knowledge base. The README states the project is for learning and discussion, and that anyone using it commercially should add their own security, compliance and licensing work.