ragflow-retrieval-mcp-server

datahwc/ragflow-retrieval-mcp-server

3.1

If you are the rightful owner of ragflow-retrieval-mcp-server and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.

The RAGFlow Retrieval MCP Server is designed to interface with the RagFlow knowledge base system, providing content-based document retrieval capabilities.

RAGFlow Retrieval MCP Server 🚀

该MCP Server用于对接RagFlow知识库系统，提供基于内容的文档检索能力。

前置要求

  • Python 3.10+
  • RagFlow

生产运行

export SERVER_HOST=http://localhost
export SERVER_PORT=80
export RAGFLOW_BASE_URL=http://localhost:80
export RAGFLOW_API_KEY=<your-api-key>
export RAGFLOW_DATASET_ID=<your-dataset-id>
export RAGFLOW_TOP_K=10
export RAGFLOW_SIMILARITY_THRESHOLD=0.2
export RAGFLOW_TIMEOUT=60

uvx ragflow-retrieval-mcp-server

开发调试

# 创建虚拟环境
uv venv
source .venv/bin/activate

# 安装依赖
uv pip install -e .

# 运行
uv run ragflow-retrieval-mcp-server

贡献指南 🤝

欢迎通过以下方式参与贡献：

  1. Fork仓库并创建特性分支
  2. 提交清晰的commit message
  3. 发起Pull Request并描述变更内容

请确保： ✅ 更新相关文档 ✅ 保持代码风格一致

许可证 📄

本项目采用

遇到问题？ 请提交 Issue 或联系 data.hwc@live.com