Gemini Mcp
Wire Google Gemini 3 into Claude Code or Cursor so your agent can generate media, run research, TTS, and sandboxed code from one MCP stdio server.
Overview
io.github.RLabs-Inc/gemini-mcp is a MCP server for the Build phase that exposes 30+ Gemini 3 tools—media, research, TTS, and code execution—to coding agents over stdio.
What is this MCP server?
- 30+ MCP tools across Gemini 3: images, video, deep research, text-to-speech, and CLI-style code execution
- stdio npm package @rlabs-inc/gemini-mcp (v0.7.1) with required GEMINI_API_KEY
- Optional GEMINI_OUTPUT_DIR for persisted generated images, video, and audio
- MCP integration server—not a SKILL.md workflow; pairs with any agent that supports MCP
- 30+ MCP tools documented in the server description
- Package version 0.7.1 (@rlabs-inc/gemini-mcp)
- Transport: stdio
Community signal: 201 GitHub stars.
What problem does it solve?
Your agent can edit repos but cannot natively call Gemini for images, video, research, speech, or sandboxed runs without custom glue code.
Who is it for?
Solo builders standardizing on Gemini 3 who want multimodal and research tools in the same MCP slot as their Claude Code or Cursor session.
Skip if: Teams that only use OpenAI or Anthropic APIs with no Gemini key, or anyone who refuses to give an agent generative-media and code-execution tool access.
What do I get? / Deliverables
After you register the server in your agent’s MCP config, the model can invoke Gemini capabilities as named tools and save outputs to GEMINI_OUTPUT_DIR when configured.
- Registered stdio MCP server in your agent config
- Agent-invokable Gemini tools for media, research, TTS, and code exec
- Generated files in GEMINI_OUTPUT_DIR when that variable is set
Recommended MCP Servers
Journey fit
Solo builders add this while assembling the agent stack—not during idea research alone—because it extends the coding agent with multimodal and execution tools. Canonical shelf is agent-tooling: the server exposes 30+ MCP tools (not a single app feature), matching how builders register LLM backends beside their repo.
How it compares
MCP integration exposing Google’s API toolkit—not an agent skill or a hosted no-code workflow.
Common Questions / FAQ
Who is io.github.RLabs-Inc/gemini-mcp for?
It is for indie developers and small teams using MCP-enabled coding agents who already have a Google Gemini API key and want one server for images, video, research, TTS, and related tools.
When should I use io.github.RLabs-Inc/gemini-mcp?
Use it during Build when you are wiring agent-tooling and need Gemini-backed media, research, or execution without writing one-off integrations for each capability.
How do I add io.github.RLabs-Inc/gemini-mcp to my agent?
Install @rlabs-inc/gemini-mcp via npm, run it as a stdio MCP server, set GEMINI_API_KEY (required) and optional GEMINI_OUTPUT_DIR, then add the server entry to your Claude Code, Cursor, or compatible client MCP configuration.