MCP Gemini
Wire Google Gemini (text, image, video, research) into your agent via MCP so builds and prototypes can call one multimodal API from the IDE.
Overview
MCP Gemini is a MCP server for the Build phase that connects your coding agent to Google Gemini for text, image, video, and research workloads.
What is this MCP server?
- MCP bridge to Google Gemini for text, image, video, and research-style tasks
- stdio transport via PyPI mcp-gemini-crunchtools (v0.3.0)
- Crunchtools-maintained server with GitHub repo mcp-gemini
- Fits multi-phase use: research summaries, prototype assets, build-time codegen assist
- MCP schema 2025-12-11 server manifest for registry discovery
- Server version 0.3.0
- PyPI identifier mcp-gemini-crunchtools
- Transport type stdio
What problem does it solve?
Your agent is locked to one provider while you need Gemini’s multimodal APIs inside the same MCP workflow you use to ship code.
Who is it for?
Solo builders standardizing on MCP who want Google Gemini multimodal calls next to their existing agent stack.
Skip if: Builders who only use Anthropic or OpenAI in-agent with no Google Cloud/API setup, or who need enterprise Gemini governance outside a personal key.
What do I get? / Deliverables
After registration, the agent can invoke Gemini-backed tools over stdio MCP during prototyping, implementation, and research without custom glue scripts.
- Agent-invokable Gemini text and multimodal operations via MCP
- Documented stdio server entry aligned with MCP registry schema
- Repeatable provider swap-in for build and prototype workflows
Recommended MCP Servers
Journey fit
Multimodal LLM access is foundational during Build when you extend the agent’s toolchain beyond a single default model. Agent-tooling is the shelf for MCP servers that add model providers and capabilities to the coding agent runtime.
How it compares
MCP model-provider integration, not a prompt library skill or hosted chat UI.
Common Questions / FAQ
Who is io.github.crunchtools/gemini for?
Developers using MCP-enabled agents who want Google Gemini text, image, video, and research capabilities inside Claude Code, Cursor, or similar tools.
When should I use io.github.crunchtools/gemini?
During build and agent-tooling setup, or when validating prototypes and running research that benefit from Gemini multimodal APIs.
How do I add io.github.crunchtools/gemini to my agent?
Install mcp-gemini-crunchtools from PyPI, add the stdio MCP server block in your agent config, and supply valid Google Gemini API credentials per the repo README.