AI MCP Service New

Oh-my-rime provides a knowledge base query service based on MCP (Model Context Protocol). With MCP, you can let AI editors (such as Cursor, Windsurf, VS Code, etc.) directly access the oh-my-rime documentation knowledge base, getting accurate configuration tutorials and usage guides during AI conversations.

In short: After configuring MCP, you can ask questions about oh-my-rime directly in your AI editor, and the AI will automatically query our knowledge base to answer you.

What is MCP?

MCP (Model Context Protocol) is an open protocol proposed by Anthropic to standardize the interaction between AI models and external data sources and tools. Think of it as a "plugin system" for AI assistants — through MCP, AI can call external tools to retrieve real-time, accurate information instead of relying solely on training data.

Service Information

Item	Description
Endpoint	https://www.mintimate.cc/mcp
Transport	Streamable HTTP
Protocol Version	2025-03-26
Available Tools	query_oh-my-rime — Semantic search of the oh-my-rime knowledge base
	get_download_links — Get client and config package download links
	get_schema_list — Get supported input schema list
	get_author_info — Get author information

Configuration in AI Editors

Cursor

Create a .cursor/mcp.json file in the project root directory, or add it in Cursor's global settings:

{
  "mcpServers": {
    "oh-my-rime-knowledge": {
      "url": "https://www.mintimate.cc/mcp",
      "transport": "streamable-http"
    }
  }
}

Windsurf

Add the following to Windsurf's MCP configuration:

{
  "mcpServers": {
    "oh-my-rime-knowledge": {
      "serverUrl": "https://www.mintimate.cc/mcp",
      "transport": "streamable-http"
    }
  }
}

VS Code (GitHub Copilot)

Create a .vscode/mcp.json file in the project root directory:

{
  "servers": {
    "oh-my-rime-knowledge": {
      "type": "http",
      "url": "https://www.mintimate.cc/mcp"
    }
  }
}

Claude Desktop

Add the following to the Claude Desktop configuration file (config file path: on macOS ~/Library/Application Support/Claude/claude_desktop_config.json, on Windows %APPDATA%\Claude\claude_desktop_config.json):

{
  "mcpServers": {
    "oh-my-rime-knowledge": {
      "url": "https://www.mintimate.cc/mcp",
      "transport": "streamable-http"
    }
  }
}

Cherry Studio

In Cherry Studio's settings, go to the MCP Servers configuration page and add a new Streamable HTTP type server:

{
  "mcpServers": {
    "oh-my-rime-knowledge": {
      "isActive": true,
      "name": "Oh My Rime Knowledge",
      "type": "streamable-http",
      "description": "Oh My Rime knowledge base query",
      "baseUrl": "https://www.mintimate.cc/mcp"
    }
  }
}

Note

Cherry Studio uses a different configuration format from other clients. Please use baseUrl instead of url, type instead of transport, and note that the name and type fields are required.

Note

The configuration format may vary slightly across different AI editors/clients. Please refer to the official documentation of your tool. The examples above are for reference only — the key point is to set the MCP endpoint https://www.mintimate.cc/mcp in the corresponding MCP settings.

Usage

Once configured, you can directly ask questions about oh-my-rime in your AI conversation, and the AI will automatically choose the appropriate tool to answer. For example:

• "How do I install oh-my-rime on macOS?" → calls query_oh-my-rime
• "How to configure fuzzy pinyin in oh-my-rime?" → calls query_oh-my-rime
• "What input schemas are available?" → calls get_schema_list
• "Give me the download links" → calls get_download_links
• "Who made Oh My Rime?" → calls get_author_info

Tool Parameters

query_oh-my-rime — Knowledge Base Semantic Search

Parameter	Type	Required	Description
query	string	✅	Natural language query, supports both Chinese and English. e.g. How to configure oh-my-rime
keyword	string	❌	Keyword filter, multiple keywords separated by semicolons. e.g. macOS;install;Rime
top_k	number	❌	Maximum number of results to return (default: 5, range: 1-10)

get_download_links — Get Download Links

Parameter	Type	Required	Description
resource	string	❌	Resource name, options: all, oh-my-rime, squirrel, weasel, fcitx5-rime, wanxiang-model, oh-my-rime-cli. Default: all

get_schema_list — Get Input Schema List

No parameters required. Returns all supported input schemas and their activation status.

get_author_info — Get Author Info

No parameters required. Returns author Mintimate's profile, blog, social media links, and open-source project information.

Note

MCP tool invocations are typically handled automatically by the AI assistant. You just need to ask in natural language — there's no need to manually pass these parameters.

Technical Implementation

This MCP service is deployed on EdgeOne Pages Cloud Functions, leveraging the vector semantic search capabilities of the CNB knowledge base. The Markdown content from the documentation repository is automatically vectorized and indexed, enabling high-precision semantic search.

The overall architecture is as follows: