Skip to main content
Glama
RAGFlow MCP Server

RAGFlow MCP Server

by AITech-Team

ragflow-mcp-server-continue MCP server

RAGFlow API MCP Server,可以查找知识库和聊天。

Components

Tools

  1. list_datasets

    • 列出所有数据集

    • 返回数据集的 ID 和名称

  2. create_chat

    • 创建一个新的聊天助手

    • 输入:

      • name: 聊天助手的名称

      • dataset_id: 数据集的 ID

    • 返回创建的聊天助手的 ID、名称和会话 ID

  3. chat

    • 与聊天助手进行对话

    • 输入:

      • session_id: 聊天助手的会话 ID

      • question: 提问内容

    • 返回聊天助手的回答

  4. retrieve

    • 检索相关信息

    • 输入:

      • dataset_ids: 数据集的 ID

      • question: 提问内容

    • 返回从知识库检索到的内容

Related MCP server: RAGFlow MCP Server

Configuration

[TODO: Add configuration details specific to your implementation]

Quickstart

Install

GitHub Copilot

.vscode/mcp.json

{
    "servers": {
        "ragflow-mcp-server": {
            "command": "uvx",
            "args": [
                "ragflow-mcp-server",
                "--api-key=ragflow-dhMzViYzJlMTM1NjExZjBiNWU5MDI0Mm",
                "--base-url=http://172.16.33.66:8060"
            ]
        }
    }
}

Continue

config.yaml

mcpServers:
  - name: RAGFlow Server
    command: uvx
    args:
      - ragflow-mcp-server
      - --api-key
      - ragflow-dhMzViYzJlMTM1NjExZjBiNWU5MDI0Mm
      - --base-url
      - http://172.16.33.66:8060

Claude Desktop

On MacOS: ~/Library/Application\ Support/Claude/claude_desktop_config.json On Windows: %APPDATA%/Claude/claude_desktop_config.json

Development

Building and Publishing

To prepare the package for distribution:

  1. Sync dependencies and update lockfile:

uv sync
  1. Build package distributions:

uv build

This will create source and wheel distributions in the dist/ directory.

  1. Publish to PyPI:

uv publish

Note: You'll need to set PyPI credentials via environment variables or command flags:

  • Token: --token or UV_PUBLISH_TOKEN

  • Or username/password: --username/UV_PUBLISH_USERNAME and --password/UV_PUBLISH_PASSWORD

Debugging

Since MCP servers run over stdio, debugging can be challenging. For the best debugging experience, we strongly recommend using the MCP Inspector.

You can launch the MCP Inspector via npm with this command:

npx @modelcontextprotocol/inspector uv --directory ragflow-mcp-server-continue run ragflow-mcp-server-continue

Upon launching, the Inspector will display a URL that you can access in your browser to begin debugging.

Available Tools

4 tools
chat聊天C

向聊天助手提问

Inputs: session_id (required), question (required).

create_chat创建聊天C

创建一个新的聊天助手,基于指定的数据集

Inputs: dataset_id (required), name.

list_datasets列出数据集B

列出 RAGFlow 中的所有数据集

Inputs: no named parameters.

retrieve检索相关信息C

从RAGFlow中指定的数据集检索相关内容

Inputs: dataset_ids (required), question (required).

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    B3.2/5.0

    Scored across 4 tools

    Disambiguation4/5

    The tools are mostly distinct with clear purposes: chat for queries, create_chat for new assistants, list_datasets for enumeration, and retrieve for content retrieval. However, chat and retrieve could potentially be confused since both involve interacting with datasets, though their descriptions clarify different intents (general Q&A vs. specific content fetching).

    Naming Consistency5/5

    All tool names follow a consistent verb-based pattern: chat, create_chat, list_datasets, and retrieve. They use simple, clear verbs without mixing conventions like camelCase or snake_case, making the naming predictable and easy to understand.

    Tool Count4/5

    With 4 tools, the count is reasonable for a RAG-focused server, covering core operations like chatting, dataset management, and retrieval. It's slightly lean but functional; adding tools for updating or deleting datasets could enhance completeness without being necessary for basic use.

    Completeness3/5

    The tools cover key RAG workflows: dataset listing, retrieval, and chat interactions. However, there are notable gaps, such as no tools for creating, updating, or deleting datasets, which limits full lifecycle management. Agents can work around this by focusing on existing datasets, but the surface is incomplete for comprehensive dataset control.

    Related MCP Connectors

    Related MCP Servers