RAGs is an open-source Streamlit application from the LlamaIndex team that lets you build a “ChatGPT over your data” entirely through natural-language conversation. Instead of configuring chunking, embeddings and vector stores by hand, you describe what you want and the app assembles the retrieval-augmented generation pipeline for you. It has earned about 6,500 stars on GitHub and is MIT licensed.
Building a first RAG system normally means a day of plumbing before you can ask your documents anything. RAGs flips the process into a chat: you upload files or point at data sources, answer a few questions, and an LLM-driven configurator selects and tunes the pipeline components — loaders, chunking strategy, embedding model, vector store and retriever settings — then generates a working Streamlit app you can chat with immediately. Because it is built on LlamaIndex, the resulting pipeline uses the framework’s mature data connectors and indexing, and everything it generates is ordinary code you can inspect and modify afterwards. The project showcases an early, influential example of using an LLM to configure other LLM systems, and remains a fast on-ramp for prototyping document chatbots.
Free. RAGs is open source under the MIT license; you only pay for the LLM API usage (for example OpenAI) that the configurator and the generated app consume.
Developers and tinkerers who want a working RAG prototype over their own documents in minutes — and a readable, generated codebase to learn from and customize afterwards.