🎙️ LexiVox

LexiVox is a fully in-browser AI-powered transcription and translation tool built using OpenAI's Whisper model served via @xenova/transformers.
It supports live mic streaming or uploading audio files (MP3) with no backend involved. All the magic runs right inside your browser.

👉 Click here to use live → LexiVox

---

🚀 Features

• 🎤 Upload .mp3 audio or record live from mic
• 💬 Transcribe in real-time using whisper-tiny.en
• 🌍 Translate audio to English
• ⚙️ Fully client-side ML using @xenova/transformers
• ✨ Clean UI with React + Vite + Tailwind CSS

---

🧠 Tech Stack

Layer	Technology
Frontend	React + Vite + Tailwind CSS
Model	openai/whisper-tiny.en
Runtime	@xenova/transformers (Web ML)
Transcription	Whisper (ASR) in browser
Deployment	Netlify

---

🗂️ Project Structure

Lexivox
├── README.md                # ← This file
├── index.html               # Entry HTML with Vite
├── vite.config.js           # Vite config
├── package.json             # Project dependencies
├── public/
│   └── vite.svg             # Static assets
├── src/
│   ├── App.jsx              # Root React component
│   ├── index.css            # Tailwind CSS config
│   ├── main.jsx             # App entry point
│   ├── assets/
│   │   └── react.svg
│   ├── components/
│   │   ├── FileDisplay.jsx     # Audio upload/view
│   │   ├── Header.jsx          # Top UI section
│   │   ├── HomePage.jsx        # Landing UI
│   │   ├── Information.jsx     # Transcription result display
│   │   ├── Transcribing.jsx    # Loading animation
│   │   ├── Transcription.jsx   # (Optional component)
│   │   └── Translation.jsx     # Translated text UI
│   └── utils/
│       ├── presets.js          # Message constants
│       ├── whisper.worker.js   # Web worker for Whisper
│       └── translate.worker.js # Web worker for translation

---

🧪 Getting Started

# 1. Clone the repo
git clone https://github.com/yourusername/lexivox.git
cd lexi-vox

# 2. Install dependencies
npm install

# 3. Start development server
npm run dev

---

🧬 Model Info

Feature	Description
Task	Automatic Speech Recognition (ASR)
Model Used	openai/whisper-tiny.en
Library	@xenova/transformers
Runtime	In-browser (CPU/WebAssembly — no backend required)
Streaming	✅ Yes — partial results are emitted in real time
Translation	✅ Optional translation available (user-triggered)
Mic Support	✅ Real-time microphone input
File Upload	✅ Accepts .mp3 and other audio formats
Dependencies	React, Vite, Xenova Transformers.js