README.md

metadata

license: apache-2.0
title: teste-teste-teste
sdk: gradio
emoji: 🏃
colorFrom: red
colorTo: red
short_description: yadayadayada
sdk_version: 5.33.0

Real-Time Screen Assistant - Premium Edition

This is a premium real-time screen assistant that integrates Google's Gemini 2.0 Live API with advanced screen recording capabilities.

Features

• 🎙️ Real-time Audio Streaming - Voice activity detection with noise filtering
• 🖥️ Professional Screen Recording - Native ScreenRecorder component with webcam overlay
• 🤖 AI Voice Responses - Bidirectional audio communication with Gemini 2.0
• 📝 Text Response Display - Real-time text responses with conversation history
• 🔄 Background Task Management - Proper async handling and cleanup
• 📊 Performance Monitoring - Real-time stats and adaptive quality

Setup

1. Set your Google AI API key:

   export GEMINI_API_KEY="your-api-key-here"
   

2. Install dependencies (automatic on HuggingFace Spaces):

   pip install -r requirements.txt
   

3. Run the application:

   python app.py
   

Components

• app.py - Main application with premium real-time integration
• gradio_screenrecorder/ - Custom Gradio component for screen recording
• requirements.txt - All necessary dependencies including custom components

Environment Variables

• GEMINI_API_KEY - Required: Your Google AI API key for Gemini 2.0 Live API

Real-time Integration

This application implements complete real-time frontend integration:

1. Continuous Audio Flow (User → Model) - Voice activity detection
2. Model Audio Output (Model → User) - AI voice responses
3. Screen Recording Integration - Professional screen capture
4. Text Response Delivery (System → User) - Real-time text display

All features are optimized for 300-second real-time sessions with adaptive quality and intelligent throttling.