Gemini SDK

The official developer toolkit for integrating Google's multimodal Gemini models into web, mobile, and backend applications.

The Gemini SDK provides direct access to Google’s most capable AI models (Pro, Flash, and Ultra) through streamlined libraries for Python, JavaScript, Go, and Swift. It handles complex multimodal inputs (text, images, video, and audio) while supporting advanced features like function calling, system instructions, and controlled JSON output. Developers use this interface to build production-ready features: automated content generation, real-time data extraction from visual assets, and sophisticated chat interfaces powered by a 2-million-token context window.

https://ai.google.dev/gemini-api/docs