Groq Audio Chat

A web-based chat application powered by Groq's LLM API with audio transcription and text-to-speech capabilities.

Features

Chat with the Llama-3.3-70b-versatile model through Groq API
Multiple ways to record audio input:
- Hold spacebar to record audio (keyboard shortcut)
- Press Cmd+Space (Mac) or Ctrl+Space (Windows/Linux) to toggle locked recording
- Hold the "Hold to Record" button
- Click the "Lock Recording" button for hands-free recording
Text-to-speech output for AI responses with multiple voice options:
- Choose from 19 different voices via dropdown
- Random voice selection option
Concise mode for shorter responses
Toggle TTS on/off

Project Structure

/
├── index.ts                 # Main entry point
├── handlers/                # API endpoint handlers
│   ├── index.ts             # Exports all handlers
│   ├── audio.ts             # Audio transcription handler
│   ├── chat.ts              # Chat completion handler
│   └── tts.ts               # Text-to-speech handler
└── ui/                      # UI components
    ├── index.ts             # Exports all UI components
    ├── html.ts              # HTML template
    ├── styles.ts            # CSS styles
    ├── tailwind.ts          # Tailwind configuration
    └── alpine.ts            # Alpine.js code

API Endpoints

GET / - Main web interface
POST /transcribe-audio - Transcribe audio recordings
POST /chat - Send chat messages to Groq LLM
POST /text-to-speech - Convert text to speech

Voice Options

The application features 19 different PlayHT voices through Groq's API:

Arista-PlayAI
Atlas-PlayAI
Basil-PlayAI
Briggs-PlayAI
Calum-PlayAI
Celeste-PlayAI
Cheyenne-PlayAI
Chip-PlayAI
Cillian-PlayAI
Deedee-PlayAI
Fritz-PlayAI
Gail-PlayAI
Indigo-PlayAI
Mamaw-PlayAI
Mason-PlayAI
Mikail-PlayAI
Mitch-PlayAI
Quinn-PlayAI
Thunder-PlayAI

Environment Variables

GROQ_API_KEY - Your Groq API key

Technologies Used

Hono - Lightweight web framework
Alpine.js - Front-end JavaScript framework
Tailwind CSS - Utility-first CSS framework
Groq API - LLM provider with Llama-3.3-70b-versatile model
Groq Whisper API - Audio transcription
PlayHT (via Groq) - Text-to-speech synthesis

Credits

This project is based on the original Val Town project by yawnxyz/groqAudioChat, with additional features and improvements including:

Refactored code structure
Enhanced recording options with keyboard shortcuts
Multiple voice selection options
Improved UI and visual feedback