gemini-mcp-tool and gemini-mcp
These are complements: the first provides CLI-based access to Gemini's processing capabilities (optimized for code analysis), while the second exposes Gemini's multimodal abilities (vision, video, chat) with integrated preview functionality, allowing assistants to leverage different aspects of Gemini depending on the task requirements.
About gemini-mcp-tool
jamubc/gemini-mcp-tool
MCP server that enables AI assistants to interact with Google Gemini CLI, leveraging Gemini's massive token window for large file analysis and codebase understanding
Implements the Model Context Protocol (MCP) to bridge Claude with Gemini CLI, exposing tools like `ask-gemini` and `sandbox-test` for safe code execution and file analysis using the `@` syntax for context injection. Integrates directly into Claude Desktop and Claude Code via stdio transport, defaulting to the `gemini-2.5-pro` model while supporting configurable model selection and sandbox mode for isolated script testing.
About gemini-mcp
houtini-ai/gemini-mcp
Google Gemini MCP - Use Google's visual, video and chat capabilities in your AI Assistant. Uses MCP Apps to preview created images, landing pages, SVGs and video
Implements 13 Gemini tools (chat with grounded search, deep research, image/video/SVG generation, image editing) via MCP stdio transport for seamless Claude Desktop integration. Grounded search is enabled by default; video generation uses Veo 3.1 with configurable resolution and audio synthesis; SVG output is production-ready vector code with style variants. Requires only a Gemini API key and `npx` one-liner for setup, or local build via npm for development environments.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work