Voxtral

A mobile-first Progressive Web App for real-time voice transcription using Mistral's Voxtral API.

Features

• 🎙️ Live Transcription - Real-time voice-to-text as you speak
• ⏱️ Recording Controls - Start, stop, pause recording
• 📝 Transcript Management - Save, edit, search, delete transcripts
• 🌍 Multi-language - 13 supported languages
• 🌙 Dark Mode - Light, dark, and system themes
• 📱 PWA - Install on any device

Tech Stack

• React 18 - UI framework
• Vite - Build tool
• Tailwind CSS - Styling
• Zustand - State management
• Dexie - IndexedDB wrapper
• Web Audio API - Audio capture
• Capacitor - Native mobile (coming soon)

Getting Started

Prerequisites

• Node.js 18+
• A Mistral API key from console.mistral.ai

Installation

# Install dependencies
npm install

# Start development server
npm run dev

Usage

1. Open the app at http://localhost:5173
2. Go to Settings and enter your Mistral API key
3. Tap the microphone button to start recording
4. Speak and watch your words appear in real-time!

Project Structure

src/
├── components/     # UI components
│   ├── ui/         # Base components (Button, Card, etc.)
│   └── *.tsx      # Feature components
├── screens/        # Page components
├── services/       # Business logic
├── stores/         # Zustand state
├── hooks/          # Custom React hooks
├── workers/        # AudioWorklet processor
├── types/          # TypeScript types
└── utils/          # Helper functions

API

This app uses the Mistral Voxtral realtime transcription API:

• WebSocket: wss://api.mistral.ai/v1/audio/transcriptions/realtime
• Model: voxtral-mini-transcribe-realtime-2602
• Pricing: $0.001/minute

License

MIT