Updated March 2026

Install Google Gemini MCP

Pick your client, copy the command, done.

Run this in a terminal.

Run in terminal

claude mcp add -e "GEMINI_API_KEY=${GEMINI_API_KEY}" google-gemini-mcp -- npx @houtini/gemini-mcp

Required:GEMINI_API_KEY+ 1 optional

Environment Variables

Set these before running Google Gemini MCP.

VariableDescriptionRequired

GEMINI_API_KEYGoogle Gemini API key from Google AI StudioYes

GEMINI_IMAGE_OUTPUT_DIROptional directory path for saving generated imagesNo

Once configured, Google Gemini MCP gives your AI agent access to:

gemini_chatChat with Google Search grounding for current information.

promptgroundingthinking_level

gemini_deep_researchRuns multiple grounded search iterations to synthesize a comprehensive report.

research_questionmax_iterationsfocus_areas

generate_imageGenerates images with optional search grounding.

promptuse_searchaspectRatio

After setup, try these prompts with your AI agent:

→What changed in the MCP spec in the last month?

→Research the current approaches to AI agent memory management and provide a summary report.

→Generate a stock price chart showing Apple (AAPL) closing prices for the last 5 trading days.

→Explain this code snippet using Gemini's reasoning capabilities.

Prerequisites & system requirements

An MCP-compatible client (Claude Code, Cursor, Windsurf, Claude Desktop, or Codex)
Node.js 18+ with npm/npx installed
GEMINI_API_KEY — Google Gemini API key from Google AI Studio

Alternative installation methods

claude mcp add -e GEMINI_API_KEY=your-api-key-here -s user gemini -- npx -y @houtini/gemini-mcp

Save the docs, env vars, and workflow around Google Gemini MCP in Conare so Claude Code, Codex, and Cursor remember it next time.