Advanced Reason MCP

Name: Advanced Reason MCP
Author: Kuon-dev

@Kuon-dev

リポジトリ

Advanced Reason MCP について

Enhanced version of "Sequential Thinking" MCP

#mcp #gemini-pro #mcp-servers #mcp-server #deepseek-r1

基本情報

カテゴリ

その他

ライセンス

MIT

ランタイム

node

トランスポート

stdio

公開者

Kuon-dev

設定

以下の設定を使って、このサーバーを MCP 対応クライアントに追加してください。

{
  "mcpServers": {
    "advanced-reflection-reason": {
      "command": "node",
      "args": [
        "/path/to/your/advanced-reason-mcp/dist/index.js"
      ],
      "env": {
        "GEMINI_API_KEY": "",
        "OPENROUTER_API_KEY": "",
        "OPENROUTER_MODEL": ""
      }
    }
  }
}

ツール

ツールは検出されませんでした

ツールは README から自動的に抽出されます。メンテナーは ## Tools という見出しの下に記載することで、このタブに反映できます。

概要

What is Advanced Reason MCP?

Advanced Reason MCP is a tool for advanced reasoning and reflection that uses the Gemini API and Open Router API. It is designed for developers who want to chain sequential thinking (Claude), Gemini critique, and Deepseek critique steps as a Model Context Protocol server.

How to use Advanced Reason MCP?

Clone the repository, install dependencies with npm install, and build using node ./bin/build.mjs. Configure the server in your claude_desktop_config.json with GEMINI_API_KEY, OPENROUTER_API_KEY, and optional OPENROUTER_MODEL (defaults to deepseek/deepseek-r1:free).

Key features of Advanced Reason MCP

Chain sequential thinking with external critique models
Integrates Gemini and Deepseek via Open Router
Configurable critique model via environment variable
Automated reasoning and reflection pipeline
Built as a lightweight MCP server

Use cases of Advanced Reason MCP

Automate multi-step reasoning and reflection tasks
Validate and critique AI-generated reasoning chains
Add external model feedback to Claude’s thinking process
Experiment with different critique models for reasoning tasks

FAQ from Advanced Reason MCP

What are the runtime requirements for Advanced Reason MCP?

Node.js 20.17 or higher is required, along with a Gemini API key and an Open Router API key.

Does Advanced Reason MCP have access to my project context?

No, the AI models do not have access to your project context.

What is the default critique model?

If OPENROUTER_MODEL is not set, it defaults to deepseek/deepseek-r1:free.

What happens if a task is too complex?

If the content becomes too long, the MCP may timeout and return no response.

How do I configure Advanced Reason MCP?

Add the server entry in claude_desktop_config.json with the node command, the path to dist/index.js, and set the GEMINI_API_KEY, OPENROUTER_API_KEY, and optionally OPENROUTER_MODEL environment variables.