Vocalization Agent

This agent synthesizes and delivers the final spoken output in a real-time translation workflow. It subscribes to a Pub/Sub topic of translated text, detects new text that is ready for speech, and uses the Google Cloud Text-to-Speech API to convert it into natural-sounding audio in the target language, selecting appropriate voices and intonation. It streams the audio back to the client application for immediate playback, or stores it in Cloud Storage or hands it to the Orchestration Agent for delivery.

The agent is part of PolyLingo, an agentic AI translator for enterprises and content creators that uses multilingual LLMs and speech models to translate across text, voice, and video. PolyLingo integrates with communication tools for live translation in conversations, meetings, and customer support, preserving tone, intent, and domain terminology for sectors such as media production, gaming, and BPOs.