This repository provides a React-based starter application for interacting with the Gemini Live API via WebSockets. It's designed for developers building applications that require real-time audio streaming, media recording (microphone, webcam, screen capture), and a unified log view for debugging. The primary benefit is a pre-built frontend structure to accelerate development with the Gemini Live API.

How It Works

The application utilizes a custom event-emitting WebSocket client to manage communication between the frontend and the Gemini API. It includes modules for handling audio input/output, capturing user media, and displaying real-time logs. A key feature demonstrated is the integration with vega-embed for rendering data visualizations based on API responses, showcasing how to define tools and functions for the Gemini model.

Quick Start & Requirements

• Install dependencies and start the development server: npm install && npm start
• Requires a Gemini API key, added to a .env file.
• Project bootstrapped with Create React App.
• Demo available at: https://gemini.google.com/app/live-api

Highlighted Details

• Real-time audio streaming and recording capabilities.
• Unified log view for development assistance.
• Example integration with vega-embed for data visualization.
• Includes example applications on other branches (e.g., demos/GenExplainer).

Maintenance & Community

This project is an experiment showcasing the Live API and is not an official Google product. Maintenance and support are provided on a best-effort basis.

Licensing & Compatibility

The repository does not explicitly state a license. The README mentions respecting copyright and trademark rights when sharing or creating derivative works.

Limitations & Caveats

As an experimental project, user mileage may vary regarding support and maintenance. The project is not an official Google product.