8000
Skip to c 7FFF ontent

Repository files navigation

title Whisper WebGPU + Llama Chat
emoji 🎙️🦙
colorFrom blue
colorTo green
sdk docker
pinned false
app_port 7860
# 🎙️ Whisper Web + Llama 3.2 🦙

ML-powered speech recognition and AI chat, right in your browser

FeaturesQuick StartModelsDevelopmentTech Stack

Report Bug | Request Feature


✨ Features

  • 🎤 Real-time Processing - Instant speech recognition with Whisper and AI responses with Llama 3.2 1B Instruct
  • 🌐 100% Client-side - No server needed, everything runs in your browser
  • WebGPU Powered - Hardware-accelerated for optimal performance
  • 🔒 Privacy First - All processing happens locally
  • 🌍 Multilingual - Supports multiple languages

🚀 Quick Start

# Clone the repository
git clone https://github.com/Reidmen/whisper-llama-web.git

# Install dependencies
pnpm install

# Start the development server
pnpm run dev

Visit http://localhost:5173 in your browser.

🤖 Models

Speech Recognition

Model Size Languages Performance
Tiny 120MB All Fast
Base 206MB All Balanced
Small 586MB All Accurate
Large V3 1.6GB All Most Accurate
Distil Small 538MB English Optimized

Chat Model

  • Llama 3.2 1B Instruct (Quantized for browser)
    • Fast local inference
    • ~1GB memory usage
    • Multi-turn conversations

💻 Development

Prerequisites

  • Modern browser with WebGPU support
  • GPU with sufficient memory
  • Microphone access (for recording)

Local Setup

  1. Clone and install dependencies
  2. Configure environment variables (if needed)
  3. Start development server
  4. Open browser at localhost:5173

Docker Deployment

docker build -t whisper-web .
docker run -p 7860:7860 whisper-web

🛠️ Tech Stack

  • Frontend: React, TailwindCSS
  • ML Framework: Transformers.js
  • Performance: WebGPU, Web Workers
  • Build Tools: Vite, TypeScript

📄 License

MIT License - see the LICENSE file for details

🙏 Acknowledgments


Built by Reidmen, inspired by the transformers.js

Releases

Sponsor this project

Packages

Contributors

Languages

0