Connects Claude to Google's Gemini API for visual analysis through two tools: analyze_image for processing image URLs (up to 16MB each) and analyze_youtube_video for extracting insights from YouTube clips without local downloads. You supply your own Gemini API key from Google AI Studio and can override the default model (gemini-flash-lite-latest) if needed. Useful when you want Claude to describe product photos, extract text from screenshots, or summarize video content, but you'd rather route the heavy lifting through Gemini's vision capabilities instead of using Claude's native image analysis. The remote version only handles URLs, though a local variant exists for filesystem access.
claude mcp add --transport http ai.smithery-artin0123-gemini-image-mcp-server https://server.smithery.ai/@Artin0123/gemini-image-mcp-server/mcp