Gemini

You run a Gemini CLI MCP Server to expose Gemini CLI prompts over MCP, allowing remote clients to send prompts and receive responses through either a local stdio server or a networked SSE server. This enables automated workflows, session management, and extension listing while keeping output clean and manageable.

How to use

Choose the transport that fits your needs. For local development and direct MCP client integration on the same machine, use the local stdio server. For remote access or deployment as a persistent service, use the SSE network server. You can resume previous Gemini sessions, list available sessions, and discover CLI extensions through the provided MCP tools.

How to install

Prerequisites you need to prepare before installing and running the server.

Install and configure Python 3.10 or higher, then set up a virtual environment and install dependencies.

cd /home/ward/MCP/gemini_mcp_server
python3 -m venv venv
source venv/bin/activate
pip install -r requirements.txt

Additional setup for SSE (network) deployment

To run the SSE server so remote clients can connect, start the SSE transport server and ensure it binds to all interfaces on port 8601.

cd /home/ward/MCP/gemini_mcp_server
source venv/bin/activate
python server_sse.py

Local stdio server setup

For a direct MCP client integration on the same machine, run the standard Python script that bundles the Gemini CLI MCP server.

python /home/ward/MCP/gemini_mcp_server/server.py

Available tools

run_gemini

Execute a Gemini CLI prompt with options such as model, output_format, and YOLO mode to auto-approve actions.

gemini_resume

Resume a previous Gemini CLI session, optionally providing a follow-up prompt.

gemini_list_sessions

List available Gemini CLI sessions for the current project.

gemini_list_extensions

List all available Gemini CLI extensions.

gemini_version

Get the Gemini CLI version information.

gemini_delete_session

Delete a Gemini CLI session by index.

gemini_json

Execute a Gemini prompt and return a structured JSON output.

gemini_with_context

Execute a Gemini prompt with specific file or directory context.

gemini_shell

Execute a shell command through Gemini CLI using the ! syntax for analysis.

gemini_chat_save

Save the current Gemini chat session.

gemini_help

Get Gemini CLI help information and available commands.