Skip to content

RAGFlow

An open-source RAG engine built on deep document understanding

Visit official site

What RAGFlow is

The engine is described as an open-source RAG engine based on deep document understanding, offering a streamlined RAG workflow for enterprises of any size. It combines a large language model for grounded question answering and supports citing evidence from data in complex formats such as Word, PPT, Excel, TXT and images.

What you can do with it

  • Extract knowledge with deep document understanding
  • Process text chunks with templates
  • Adjust visual text chunks by hand
  • Configure the large language model and embedding model

Application scenarios

  • Run automated question answering over enterprise documents
  • Analyse legal files and produce summaries automatically
  • Organise teaching materials and answer questions in education

Who it is for

  • Enterprise users who handle large volumes of documents and data
  • Researchers who need useful information from complex data
  • Business processes that automate document handling and report generation

Getting started

  • Make sure the system meets the CPU, RAM, disk and Docker minimums
  • Adjust vm.max_map_count for the Docker requirement
  • Clone the RAGFlow GitHub repository locally
  • Build the prebuilt Docker image and start the server

Pros & cons

✓ What we like

  • Cited answers can be traced back to the source data
  • Word, PPT, Excel, TXT and image sources are supported

! What to watch out for

  • The documented setup requires Docker and a raised vm.max_map_count value
  • Chunking may need manual intervention through the visual interface

FAQ

Which data sources are supported?

The feature list names Word, PPT, Excel, TXT and images.

What does setup require?

The steps list CPU, RAM, disk and Docker minimums, adjusting vm.max_map_count, cloning the repository and starting a Docker image.

Last reviewed: 2026-11-15

More AI office assistant tools

View all →

How we review