GitHub radar

RAGFlow: Open-Source AI Chatbot from Your Documents

RAGFlow is an open-source RAG engine from infiniflow: upload your PDFs, sync Notion or Google Drive, and get a working AI chatbot that answers from your own knowledge base.

RAGFlow is an open-source RAG engine from infiniflow, released in late 2023, with nearly 91,000 stars. It provides a full pipeline for building AI assistants that answer from your own documents: upload PDFs, Word files, or sync from Confluence, S3, Notion, Discord, and Google Drive. The engine handles document parsing (including MinerU and Docling), chunking, embedding, and retrieval. It includes a visual agentic workflow builder and pre-built agent templates. Answers can be delivered via a built-in chat UI or through integrations with Telegram, Discord, Feishu, and Line. Supports any LLM via API, including OpenAI, DeepSeek, and Gemini. MCP support was added in mid-2025. A cloud version is available at cloud.ragflow.io. Self-hosting requires CPU ≥ 4 cores, RAM ≥ 16 GB, and Docker.

Why a vibe-coder should care

If you need an AI assistant that reliably answers from your own documents — product docs, notes, or research materials — without hallucinating, RAGFlow is one of the most mature open-source options available. The cloud version lets you test before committing to a local deployment.

How to install

Copy this and send it to your agent — Claude Code, Codex, any of them:

Deploy RAGFlow from https://github.com/infiniflow/ragflow — follow the README (docker compose up), run it locally, open the browser UI, and show me the interface. Ask me if you need an LLM API key.

Self-hosting needs a machine with 16+ GB RAM and Docker. To try without installing, use cloud.ragflow.io.

Open on GitHub