Gemini Audio Mcp

Wraps Google's Gemini 2.5 Live API and DeepMind's Lyria 3 models for audio synthesis tasks. Exposes four main tools: generate_soundscape for ambient textures, generate_music for structured compositions with BPM and key control, generate_voice for narration using native audio, and transition_soundscape for crossfading between environments. Uses a hybrid architecture with WebSocket connections for voice and REST calls for music generation, plus an internal Rust pipeline that handles PCM looping to extend short clips into seamless audio. Requires FFmpeg for transcoding and a Google AI Studio API key with Lyria access. Reach for this when you need programmatic audio generation in Claude workflows, like building soundscapes for games, generating background music, or creating dynamic narration.