RAGFlow
RAGFlow is an open-source Retrieval-Augmented Generation (RAG) engine and orchestration platform developed by the InfiniFlow team. It provides developers with end-to-end RAG pipeline management capabilities based on document parsing, knowledge base construction, and large language model integration. The platform supports deep document understanding across multiple file formats and offers a visual workflow orchestration interface. It is widely used in enterprise knowledge base question answering, intelligent customer service, and document retrieval applications.
Core Facts
RAGFlow is an open-source RAG engine that provides complete knowledge base management and intelligent question-answering functionality.
RAGFlow depends on multiple services including Elasticsearch (full-text search), Redis (caching), and MySQL (metadata storage)
RAGFlow is deployed based on Docker containerization and runs five containers upon startup: Redis, MinIO, MySQL, Elasticsearch, and RAGFlow Server.
Timeline (last 90 days)
The tutorial explicitly avoids deploying RAGFlow using Docker Desktop on Windows, on the grounds that this type of installation will have various hard-to-troubleshoot problems.
The core functions of the RAGFlow main interface are divided into four major modules: knowledge base, chat, search, and Agent, plus a file management entry.
RAGFlow's implementation of MCP differs from the conventionally understood MCP functionality, so before integration one should consult the official documentation to confirm its role, transport method, and available tool list.
RAGFlow provides two calling methods: an HTTP API and a Python SDK.
Chat applications or Agents developed in RAGFlow can be embedded into websites.
In RAGFlow, the vector dimension of the embedding model must match the dimension selected when the knowledge base was created, otherwise retrieval will fail.
When connecting Ollama in RAGFlow, embedding models (such as bge-m3 and nomic-embed-text) are used for vectorizing documents and queries, while chat models (such as qwen2 and llama3) are used to generate answers, and the two must be added separately.
RAGFlow supports connecting to the Ollama inference framework, allowing local models to be used for knowledge base question answering, with both embedding models and generation models able to run locally.
Compared with platforms such as Dify, RAGFlow has fewer Agent nodes and lacks advanced control structures such as loops and subflows.
RAGFlow's Agent workflow consists of several predefined nodes, with common types including the start node, knowledge base retrieval node, LLM generation node, conditional branch node, code execution node, and end node.
16 more timeline events
All Facts (20)
RAGFlow is an open-source RAG engine that provides complete knowledge base management and intelligent question-answering functionality.
80%VerifiedRAGFlow depends on multiple services including Elasticsearch (full-text search), Redis (caching), and MySQL (metadata storage)
65%VerifiedRAGFlow is deployed based on Docker containerization and runs five containers upon startup: Redis, MinIO, MySQL, Elasticsearch, and RAGFlow Server.
65%Unverified53AI Hub支持与Coze、Dify、FastGPT、RAGFlow等主流AI开发平台无缝集成
90%UnverifiedRAGFlow集成了AI Agent功能，提供可视化画布和预置模板，支持客服、数据分析等典型场景
90%UnverifiedRAGFlow的聊天助理提供可溯源的回答机制，用户可以查看AI回答具体引用了哪些知识库原文
90%UnverifiedRAGFlow由Infiniflow团队开发，专注于RAG流程的深度优化，采用了DeepDoc等自研文档理解技术
85%Unverified在RAG框架市场中，LangChain和LlamaIndex作为编排层框架，Dify、FastGPT、RAGFlow是面向业务用户的完整平台产品
85%UnverifiedOllama的OLLAMA_HOST环境变量需设置为0.0.0.0:11434，以便Docker容器中的RAGFlow能访问本机的Ollama服务
85%UnverifiedOllama listens only on localhost by default; the environment variable OLLAMA_HOST must be configured to 0.0.0.0:11434 to listen on all IPs, making it accessible to RAGFlow inside a virtual machine.
60%UnverifiedWhen adding a local DeepSeek in RAGFlow, the model type must be selected as chat, and no API key is required for local deployment.
60%UnverifiedRAGFlow supports a parent-child chunking strategy — using small chunks for retrieval matching while returning parent chunks containing more context to the LLM
60%UnverifiedRAGFlow integrates the OCR capabilities of the DeepDoc project, enabling it to recognize text content in scanned PDFs
60%UnverifiedRAGFlow支持多种文档格式，包括PDF、Word、Excel等
60%UnverifiedRAGFlow employs Deep Document Understanding technology, which can recognize layout information such as heading hierarchy, table structure, and headers and footers within documents.
60%UnverifiedThe tutorial explicitly avoids deploying RAGFlow using Docker Desktop on Windows, on the grounds that this type of installation will have various hard-to-troubleshoot problems.
50%UnverifiedRAGFlow's Agent module uses workflow-style orchestration, similar in form to the visual flows in Dify or Coze.
50%UnverifiedThe number of Agent nodes currently supported by RAGFlow is relatively limited.
50%UnverifiedRAGFlow's Agent workflow consists of several predefined nodes, with common types including the start node, knowledge base retrieval node, LLM generation node, conditional branch node, code execution node, and end node.
50%UnverifiedCompared with platforms such as Dify, RAGFlow has fewer Agent nodes and lacks advanced control structures such as loops and subflows.
50%