Gemini CLI MCP Server

Gemini CLI MCP Server

centminmod/gemini-cli-mcp-server is a real MCP project that bridges Google's Gemini CLI into MCP-compatible clients such as Claude Code and Claude Desktop.

What It Provides

The server exposes Gemini-powered tools through the Model Context Protocol:

Tool	Description
gemini_prompt	Send a prompt to Gemini and return the response
gemini_code_review	Review code with Gemini and return structured feedback
gemini_plan_evaluation	Evaluate an implementation plan for feasibility and gaps
gemini_solution_verify	Verify a proposed solution against requirements
gemini_collaborate	Run a multi-turn collaboration session via Gemini

Prerequisites

Before running the MCP server, you need:

1. Gemini CLI installed — follow the Gemini CLI setup guide
2. Google API key — set GEMINI_API_KEY or authenticate via gcloud
3. OpenRouter API key (optional) — for multi-model routing through OPENROUTER_API_KEY
4. Node.js 18+ for running the MCP server

Installation

git clone https://github.com/centminmod/gemini-cli-mcp-server
cd gemini-cli-mcp-server
npm install
npm run build

Configuration in Claude Code

Add to your .mcp.json or Claude Code settings:

{
  "mcpServers": {
    "gemini-cli": {
      "command": "node",
      "args": ["/path/to/gemini-cli-mcp-server/dist/index.js"],
      "env": {
        "GEMINI_API_KEY": "your-api-key"
      }
    }
  }
}

Use Cases

Code review with a second opinion: Claude Code implements a feature, then calls gemini_code_review to get Gemini's perspective before marking the task complete.

Plan validation: Before executing a multi-file refactor, use gemini_plan_evaluation to identify risks and missing steps.

Multi-model collaboration: Route complex reasoning tasks to Gemini while keeping code execution in Claude, combining the strengths of both models.

When to Choose This MCP

Choose Gemini CLI MCP Server when:

• You want Claude Code to delegate analysis or code review to Gemini
• You are building multi-model workflows that leverage both Anthropic and Google models
• You need a structured MCP interface to Gemini's capabilities rather than raw API calls

Alternatives

If you need Gemini integration without the CLI dependency, consider OpenRouter-based setups that route to Gemini models directly through a single API endpoint. Browse aimcp.info for related MCP servers.