Gemini Live API

概要

Empower Claude to build sophisticated voice-enabled applications using the Google Gemini Live API. This skill provides comprehensive technical guidance on managing low-latency WebSocket connections, handling natural speech interruptions, and configuring bidirectional audio streaming for human-like interactions. It covers complex patterns including tool integration through function calling, session persistence with resumption tokens, and voice customization for specific tones and cadences. Whether you are debugging audio sample rate conversions or implementing secure ephemeral token authentication for production, this skill serves as an expert resource for high-performance voice AI development.

主な機能

• Secure session management including resumption tokens and context compression
• Advanced Voice Activity Detection (VAD) for natural conversation and interruption handling
• Tool integration and function calling with asynchronous execution support
• Bidirectional audio streaming with real-time processing and sample rate management
• Voice customization for adjusting cadence, tone, and multi-language support
• 3 GitHub stars

ユースケース

• Building low-latency voice agents that can handle natural user interruptions
• Developing real-time customer support bots with integrated function calling for backend tasks
• Implementing voice-controlled applications with customized speech characteristics and prosody