Team Ai
Apppublic

sam-codecub/voice-ai-assistant

sourceHugging Faceupdated 10mo agoView on Hugging Face
0likes
App README

Voice AI Assistant

A voice-enabled AI chatbot built with Streamlit and Google Gemini API, featuring automatic speech-to-text transcription and multi-language support.

🚀 Live Deployments

This project is deployed on multiple platforms:

Try the live demo on Hugging Face Spaces!

Features

  • —Voice Input: Click-to-record voice input with automatic transcription
  • —Auto-Send: Messages are automatically sent after voice transcription
  • —Multi-Language Support: Supports 12 languages including English, Spanish, French, German, Italian, Portuguese, Chinese, Japanese, Korean, Hindi, and Arabic
  • —Voice Commands: Control the app with voice commands like "clear chat" or "change personality"
  • —Custom Personalities: Currently features a Clash Royale themed AI personality
  • —Chat History: Full conversation history with message display
  • —Manual Text Input: Type messages manually with a text area and send button

Technologies Used

  • —Streamlit: Web interface framework
  • —Google Gemini API: AI language model (gemini-2.5-flash)
  • —SpeechRecognition: Speech-to-text conversion using Google's speech recognition
  • —audio-recorder-streamlit: Browser-based audio recording component
  • —python-dotenv: Environment variable management

Installation

  1. 1.Clone the repository:
bash
git clone https://github.com/CodeCubCA/voice-ai-assistant-samCodeCub.git
cd voice-ai-assistant-samCodeCub
  1. 1.Install dependencies:
bash
pip3 install -r requirements.txt
  1. 1.Set up your environment variables:
  2. 2.Copy .env.example to .env
  3. 3.Add your Google Gemini API key to .env:
GEMINI_API_KEY=your_api_key_here
  1. 1.Run the application:
bash
streamlit run app.py

Usage

Voice Input

  1. 1.Click the microphone button
  2. 2.Speak clearly when recording
  3. 3.Click again to stop recording
  4. 4.Your message will be automatically transcribed and sent

Manual Input

  1. 1.Type your message in the text area
  2. 2.Click the "Send" button

Voice Commands

Speak these commands to control the app:

  • —Clear Chat: "Clear chat", "Clear conversation", or "Delete history"
  • —Change Personality: "Change personality to Clash Royale"

Configuration

Language Selection

Choose your preferred language for voice recognition from the sidebar. Supported languages:

  • —English (US/UK)
  • —Spanish
  • —French
  • —German
  • —Italian
  • —Portuguese
  • —Chinese (Mandarin)
  • —Japanese
  • —Korean
  • —Hindi
  • —Arabic

Personality Selection

Select different AI personalities from the sidebar (currently Clash Royale themed).

Tips for Better Voice Recognition

  • —Speak in a quiet environment
  • —Be close to your microphone
  • —Speak clearly at normal pace
  • —Use short sentences
  • —Grant microphone permissions in your browser

Project Structure

voice-ai-assistant/
├── app.py                 # Main application file
├── requirements.txt       # Python dependencies
├── .env                   # Environment variables (not in git)
├── .env.example          # Environment variables template
├── .gitignore            # Git ignore file
├── README.md             # This file
└── test files/           # Testing utilities
    ├── test_mic.html
    ├── test_audio_simple.py
    └── test_voice.py

Requirements

  • —Python 3.7+
  • —Modern web browser with microphone support
  • —Internet connection for speech recognition and AI API
  • —Google Gemini API key

Troubleshooting

Voice input not working:

  • —Check browser microphone permissions
  • —Try the test_mic.html file to verify microphone access
  • —Ensure no other app is using your microphone

Transcription errors:

  • —Speak more clearly and slowly
  • —Reduce background noise
  • —Try a different language setting
  • —Check your internet connection

API errors:

  • —Verify your Gemini API key in .env
  • —Check API quota limits
  • —Ensure internet connection is stable

License

This project is for educational purposes.

Credits

Built with Streamlit and Google Gemini API