GCP Architect Skill

This skill defines the "Cutting-Edge Architect" patterns for building advanced multimodal agents on Google Cloud Platform. It integrates the latest experimental models for Vision, Video, and Audio.

Tier 4: Audio Cortex (Lyria RealTime)

Model: models/lyria-realtime-exp Protocol: WebSocket (Bidirectional Streaming) Output: Raw 16-bit PCM Audio @ 48kHz, Stereo

1. Connection & Streaming (Python)

Use the google.genai SDK with asyncio. The connection must be handled asynchronously.

import asyncio
from google import genai
from google.genai import types