Gemini MCP Server Architecture Plan

Name: Gemini MCP Server Architecture Plan
Author: ivangrynenko

@ivangrynenko

Repository

About Gemini MCP Server Architecture Plan

MCP server for orchestrating multiple Gemini AI agents with persistent sessions

#gemini #mcp #ai-agents

Basic information

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "gemini-mcp-ivangrynenko": {
      "command": "python",
      "args": [
        "-m",
        "gemini_mcp"
      ]
    }
  }
}

Tools

No tools detected

We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.

Overview

What is Gemini MCP Server Architecture Plan?

An MCP server that enables Claude to orchestrate multiple Gemini AI agents dynamically, with each agent maintaining its own conversational context through persistent sessions. It is designed for developers using Claude Desktop to create, manage, and hand off context between specialized AI agents in real time.

How to use Gemini MCP Server Architecture Plan?

Install with pip install -r requirements.txt, set your GEMINI_API_KEY in a .env file, and add the server configuration to Claude Desktop’s MCP settings under the name gemini-orchestrator. Run the server with python -m gemini_mcp and then invoke tools like ai_agents_create, ai_agents_send_message, and ai_agents_handoff from Claude.

Key features of Gemini MCP Server Architecture Plan

Dynamic agent creation on demand during conversations.
Six MCP tool functions for full agent lifecycle.
Context handoff between agents with optional full history.
In-memory (RAM) or persistent SQLite storage.
Automatic session cleanup and error handling.
Full conversation history tracking per agent.

Use cases of Gemini MCP Server Architecture Plan

Setting up a development team of agents (e.g., business analyst, architect) for a project.
Gathering requirements from one agent and passing them to another for technical design.
Managing multiple short-lived agents for different sub-tasks within a single Claude session.

FAQ from Gemini MCP Server Architecture Plan

What is the purpose of this server?

It allows Claude to create and interact with multiple Gemini AI agents during a conversation, each with its own role and context, and hand off information between them.

How do I install and configure the server?

Clone the repository, install dependencies, create a .env file with your Gemini API key, and add a gemini-orchestrator entry to Claude Desktop’s MCP configuration JSON.

What storage options are available?

Two options: in‑memory (fast, data lost on restart) for development/testing, and SQLite (persistent, slightly slower) for production.

Can I list all active agents?

Yes, use the ai_agents_list tool to get a list of all agents with their roles, session IDs, creation timestamps, and message counts.

What should I do if I get a “GEMINI_API_KEY not found” error?

Ensure you have set the environment variable in your .env file and that the file is correctly loaded. Also verify the key is valid and that you’re using the correct casing.