RAGFlow provides a RAG workflow centered on deep document understanding and grounded citations for large language models. Teams can configure their own LLMs and embedding models, but the Docker image relies on external LLM and embedding API services.
Project overview
The project focuses on quality document chunking from complex, unstructured formats with the goal of reducing hallucinations and providing traceable citations.
Project type
AI Agent · RAG · Document Processing
Use cases
Knowledge Q&A
Deployment
Docker / Docker Compose
License
Apache-2.0
Best for
Developers, data teams, and enterprise teams who need to process complicated unstructured documents and require grounded citations for LLM outputs.
Key capabilities
Performs knowledge extraction from unstructured data with complicated formats.
Supports data formats including Word, slides, excel, txt, images, scanned copies, structured data, and web pages.
Limitations and risks
All Docker images are built for x86 platforms; Docker images for ARM64 are not currently offered.
This image relies on external LLM and embedding API services, which are required for inference and incur associated costs.
Getting started
Clone the repository, navigate to the docker directory, check out a specific version tag, set vm.max_map_count to at least 262144, and start the server with docker compose. Access requires logging in via a web browser. The setup also requires configuring LLM API keys.
Alternatives and comparisons
Uses a hierarchical tree index and LLM reasoning for retrieval without vector databases or chunking, contrasting with the deep document chunking approach.
Ingests unstructured documents and outputs persistent, navigable memory by reconstructing the document hierarchy.
Full-stack platform with multi-provider agents, Qdrant-backed RAG, role-based admin dashboard, and an embeddable streaming chat widget.
GitHub project description: RAGFlow is a leading open-source Retrieval-Augmented Generation (RAG) engine that fuses cutting-edge RAG with Agent capabilities to create a superior context layer for LLMs
README: RAGFlow enables developers to transform complex data into high-fidelity, production-ready AI systems with exceptional efficiency and precision.
README: Deep document understanding-based knowledge extraction from unstructured data with complicated formats.
README: Streamlined RAG orchestration catered to both personal and large businesses.
README: Quick view of the key references and traceable citations to support grounded answers.