Building a Multilingual Live Translation App with Gemini API and Google Cloud Run

In this video, we explore the new Gemini 3.5 live translation model available via the Gemini API. We demonstrate how to build an application that broadcasts speech in multiple languages using Gemini live translate, deployed on Google Cloud Run with Next.js.

■ Application Architecture
- Using Gemini API with Life Kit and Google Cloud Run
- Setting up live translate sessions per requested language
- Managing listeners and closing sessions when the last listener leaves

■ Key Features
- Long-running web sockets enabled by deploying Next.js to Google Cloud Run
- WebRTC for audio broadcasting and data channels for captions
- Open-source demo available on the Google Gemini GitHub account

This video is ideal for developers interested in real-time translation, WebRTC, and cloud deployment. You will learn how to integrate the Gemini API, manage live sessions, and leverage Cloud Run for scalable, long-lived connections.