AI Voice Trainer

A voice-enabled workout timer application that uses ElevenLabs voice agents and Gemini AI to create and run personalized workout routines with audio guidance.

Features

• Voice conversation with ElevenLabs agent to create custom workouts
• AI-generated workout plans using Google's Gemini API
• Voice-guided workout routines with timers and exercise descriptions
• Customizable workout settings (type, duration, intensity)
• Responsive design with shadcn/ui components
• Exercise form cues and motivational messages

Prerequisites

• Node.js (v16 or higher)
• npm package manager
• ElevenLabs API key (get one at ElevenLabs)
• ElevenLabs voice agent ID (create one at ElevenLabs Voice Agents)
• Google Gemini API key (get one at Google AI Studio)

Setup

1. Clone the repository:

git clone https://github.com/GitlyHallows/ai-voice-trainer.git
cd ai-voice-trainer

2. Install dependencies:

npm install

3. Configure environment variables:
   • Copy .env.example to .env:

     cp .env.example .env

   • Edit .env and add your ElevenLabs and Gemini API keys:

     VITE_ELEVENLABS_API_KEY=your_elevenlabs_api_key_here
     VITE_GEMINI_API_KEY=your_gemini_api_key_here

Run the server

Run the server:

npm run dev

The application will be available at http://localhost:8080

Using the Application

1. Voice Selection:

   • Enter your ElevenLabs API key if not already set in .env
   • Select a voice from the dropdown to use for the workout audio

2. Creating a Workout:

   • Click the microphone button in the bottom-right corner to start a conversation
   • Tell the voice agent what kind of workout you want (e.g., "I want a 25-minute leg workout")
   • The agent will use Gemini AI to generate a personalized workout plan
   • When you're ready to start, tell the agent "I'm ready to begin" or manually disconnect

3. Workout Player:

   • After the workout is generated, use the play button to start the workout
   • The player will automatically guide you through each exercise with voice instructions
   • You can use the controls if you'd like to pause, resume, or skip to the next exercise/phase

Project Structure

src/
├── components/     # React components including VoiceChat and WorkoutPlayer
├── contexts/      # React contexts
├── data/          # Sample workout data
├── hooks/         # Custom React hooks including useWorkoutAudio
├── lib/           # Utility functions including workout parser and generator
└── pages/         # Application pages

The built files will be in the dist directory.

Technologies Used

• Vite + React + TypeScript
• shadcn-ui + Tailwind CSS
• ElevenLabs API for voice synthesis and agent conversation
• Google Gemini API for workout plan generation
• Web Speech API for audio playback

Deployment

You can deploy this project using platforms like Netlify, Vercel, or GitHub Pages.

License

MIT

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.