RAGFlow
RAG engine built on deep document understanding.
- Star 数
- 92.3K
- Fork 数
- 11K
- 未关闭 Issue
- 1.4K
- 最近提交
- 2026-10-15
数据来自 GitHub API · 更新于 2026-10-15
RAGFlow 是什么？
RAGFlow focuses on high-quality retrieval from complex documents — PDFs, tables, scans — using layout-aware parsing and chunking templates, with grounded citations in answers and an agent builder.
主要功能
- Deep document parsing (layout, tables, OCR)
- Template-based chunking
- Citations with traceable sources
- Agent workflows
系统要求
- 平台
- Docker · Linux
- 内存
- 16 GB+
- 磁盘
- 50 GB+
- 备注
- CPU with 4+ cores; set vm.max_map_count ≥ 262144.
如何安装 RAGFlow
根据项目 README 整理。开源项目更新很快，请以 GitHub 页面上的最新说明为准。
Docker Compose
- 1Clone and start.
git clone https://github.com/infiniflow/ragflow.git cd ragflow/docker docker compose -f docker-compose.yml up -d - 2Open http://localhost and configure your LLM provider.
优势
- Best-in-class document parsing
- Apache license
不足
- Heavy resource requirements
RAGFlow 替代品对比
| 软件 | Star 数 | 许可证 | 平台 | 显存 |
|---|---|---|---|---|
| RAGFlow RAG engine built on deep document understanding. | 92.3K | Apache-2.0 | — | |
| AnythingLLM All-in-one desktop and Docker app for private document chat and agents. | 67.1K | MIT | — | |
| Dify Visual platform for building LLM apps, RAG pipelines and agents. | 158.5K | Dify Open Source License (Apache-2.0 based) | — | |
| Open WebUI Self-hosted ChatGPT-style interface for Ollama and OpenAI-compatible APIs. | 155.7K | Open WebUI License (BSD-3-Clause based) | — | |
| LibreChat Open-source ChatGPT clone supporting every major AI provider. | 45.7K | MIT | — | |
| LobeChat Modern AI chat framework with plugins, knowledge base and agent market. | 83.2K | LobeHub Community License | — |
常见问题
RAGFlow 免费吗？
是的——RAGFlow 是开源软件，采用 Apache-2.0 许可证。商用前请确认软件及模型权重的许可条款。
RAGFlow 支持 Windows、Mac 和 Linux 吗？
RAGFlow 支持：Docker, Linux。
运行 RAGFlow 需要什么配置？
RAM: 16 GB+
如何安装 RAGFlow？
1. Clone and start. 2. Open http://localhost and configure your LLM provider.
RAGFlow 有哪些好的替代品？
常见替代品包括 AnythingLLM, Dify, Open WebUI, LibreChat，详见本页对比表。