Google Gemini API Integration

About

This skill provides a comprehensive implementation guide and boilerplate for the Google Gemini API, ensuring developers use the current @google/genai SDK instead of deprecated versions. It supports advanced features like Gemini 3 preview, multimodal inputs including video and PDFs, function calling with parallel execution, and Thinking Mode for complex reasoning. Designed for both Node.js environments and edge runtimes like Cloudflare Workers, it streamlines the process of building sophisticated AI applications with 1M+ token context windows, cost-effective context caching, and real-time Google Search grounding.

Key Features

• 91 GitHub stars
• Advanced Thinking Mode configuration for complex reasoning
• Full multimodal support for images, video, audio, and PDFs
• Latest SDK implementation using @google/genai v1.30.0+
• Dual-path architecture for Node.js and Fetch-based edge runtimes
• Function calling with parallel and compositional execution

Use Cases

• Building AI agents with real-time web grounding and tool use
• Migrating legacy Gemini integrations to the modern SDK architecture
• Processing large-scale multimodal documents and high-volume video files