Gemini
Attach Google Gemini chat, research, and grounding to your MCP agent stack during build and across research tasks.
Overview
Gemini is a MCP server for the Build phase that exposes Google Gemini AI chat, research, and grounding to your coding agent.
What is this MCP server?
- MCP server for Google Gemini with chat, research, and grounding surfaces
- Requires GEMINI_API_KEY from Google AI Studio
- npm @houtini/gemini-mcp at version 1.4.2 with stdio transport
- Lets one agent orchestrate Gemini alongside Claude or Cursor models
- Supports grounded answers for research-heavy solo-builder workflows
- Server version 1.4.2
- npm package @houtini/gemini-mcp
- Required secret env GEMINI_API_KEY
Community signal: 25 GitHub stars.
What problem does it solve?
Switching between your IDE agent and a separate Gemini tab breaks context when you need grounded research or a second model opinion.
Who is it for?
Builders who standardize on MCP and want Google Gemini as an alternate or specialist model in Claude Code or Cursor.
Skip if: Teams forbidden from sending repo context to Google APIs or who only need financial or SEO-specific Houtini servers.
What do I get? / Deliverables
Gemini runs as MCP tools inside your agent so chat, research, and grounding stay on the same thread as your code and specs.
- Gemini chat and research responses inside MCP tool results
- Grounded summaries tied to agent prompts
- Optional second-model perspective on specs and prototypes
Recommended MCP Servers
Journey fit
Gemini is wired in during build as the canonical integrations shelf for a second model provider in the agent toolchain. MCP registration and API keys are integration work that lands in build/integrations even when you use Gemini earlier for research.
How it compares
Google LLM MCP integration, not a replacement for your primary code-generation agent skill.
Common Questions / FAQ
Who is Gemini for?
Solo builders using MCP who want Google Gemini for chat, research, or grounding without leaving Claude Code, Cursor, or a compatible host.
When should I use Gemini?
Use it when you need Gemini-specific grounding or multimodal research inside the same agent session as implementation or validation work.
How do I add Gemini to my agent?
Obtain a GEMINI_API_KEY from Google AI Studio, install @houtini/gemini-mcp, configure stdio MCP in your client, and restart so Gemini tools appear.