Skip to content

Repository files navigation

AudioKeep - Professional AI Audio Restoration Platform

AudioKeep is a world-class AI-powered audio restoration and enhancement platform designed for professionals in archival preservation, audio forensics, broadcast restoration, and historical audio digitization.

Features

Audio Restoration

  • Advanced noise reduction (background noise, hiss, hum, buzz)
  • Click and pop removal (vinyl/tape damage)
  • Spectral repair for damaged audio sections
  • Declipping for distorted audio restoration
  • Electrical interference removal (50/60Hz hum)

Audio Enhancement

  • AI-powered speech enhancement
  • Audio super-resolution (upsample to 48kHz/96kHz/192kHz)
  • Voice isolation and source separation
  • Bandwidth extension
  • Dynamic range enhancement

Professional Features

  • Audio forensics tools for law enforcement
  • Detailed spectral analysis
  • Authentication and tampering detection
  • Forensic reporting
  • Batch processing capabilities

User Features

  • Real-time A/B comparison player
  • Waveform and spectrogram visualization
  • Credit-based pricing model
  • Multiple quality tiers
  • API access for enterprise

Technology Stack

Backend

  • FastAPI - Modern Python web framework
  • PyTorch - AI model inference
  • Celery - Distributed task queue
  • PostgreSQL - Primary database
  • Redis - Queue and caching
  • MinIO/S3 - File storage

Frontend

  • React 18+ with TypeScript
  • Vite - Build tool
  • TailwindCSS - Styling
  • Shadcn/ui - Component library
  • React Query - Data fetching
  • Zustand - State management
  • Wavesurfer.js - Audio visualization

AI Models

  • Resemble Enhance - Speech denoising and enhancement
  • AudioSR - Audio super-resolution
  • DeepFilterNet - Advanced noise reduction
  • Demucs v4 - Source separation
  • FlashSR - Fast super-resolution

Hardware Requirements

Development

  • 8GB+ RAM
  • 4+ CPU cores
  • 10GB disk space

Production (Recommended)

  • GPU: NVIDIA RTX 6000 Ada (48GB VRAM)
  • CPU: AMD EPYC 9354 (32 cores, 64 threads)
  • RAM: 128GB
  • Storage: 1TB NVMe SSD

Quick Start

Prerequisites

  • Docker and Docker Compose
  • Node.js 18+ (for frontend development)
  • Python 3.11+ (for backend development)
  • NVIDIA GPU with CUDA 12.x (for production)

Development Setup

  1. Clone the repository
git clone https://github.com/yourusername/audiokeep.git
cd audiokeep
  1. Set up environment variables
cp .env.example .env
# Edit .env with your configuration
  1. Start development environment
docker-compose -f docker-compose.dev.yml up
  1. Access the application

Production Deployment

See DEPLOYMENT.md for detailed production deployment instructions.

Project Structure

audiokeep/
├── backend/
│   ├── app/
│   │   ├── api/v1/endpoints/    # API route handlers
│   │   ├── core/                # Core configuration
│   │   ├── models/              # Database models
│   │   ├── schemas/             # Pydantic schemas
│   │   ├── services/            # Business logic
│   │   ├── ai_models/           # AI model integrations
│   │   └── db/                  # Database utilities
│   ├── tests/                   # Backend tests
│   └── requirements.txt
├── frontend/
│   ├── src/
│   │   ├── components/          # React components
│   │   ├── pages/               # Page components
│   │   ├── hooks/               # Custom React hooks
│   │   ├── services/            # API services
│   │   ├── store/               # State management
│   │   └── types/               # TypeScript types
│   ├── public/                  # Static assets
│   └── package.json
├── deployment/
│   ├── docker/                  # Dockerfiles
│   └── kubernetes/              # K8s manifests
├── docs/                        # Documentation
├── scripts/                     # Utility scripts
├── ARCHITECTURE.md              # System architecture
└── docker-compose.yml

API Documentation

Once the backend is running, visit http://localhost:8000/docs for interactive API documentation (Swagger UI).

Credit System

AudioKeep uses a credit-based pricing model:

Free Tier

  • 50 credits on signup
  • Standard quality processing
  • Max 100MB file size
  • Max 30 minutes duration

Credit Packs

  • Starter: 100 credits - $9.99
  • Professional: 500 credits - $39.99
  • Studio: 1,500 credits - $99.99
  • Enterprise: 5,000 credits - $299.99

Monthly Subscriptions

  • Pro: $29.99/month (500 credits)
  • Studio: $79.99/month (1,500 credits)
  • Forensic: $199.99/month (4,000 credits + forensic tools)

Credit Usage Examples

  • 1-minute podcast cleanup: ~2 credits
  • 10-minute interview restoration: ~35 credits
  • 60-minute archival tape (ultra quality): ~240 credits
  • 5-minute forensic analysis: ~30 credits

Development

Backend Development

cd backend
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate
pip install -r requirements.txt
uvicorn app.main:app --reload --host 0.0.0.0 --port 8000

Frontend Development

cd frontend
npm install
npm run dev

Running Tests

# Backend tests
cd backend
pytest

# Frontend tests
cd frontend
npm test

Contributing

We welcome contributions! Please see CONTRIBUTING.md for guidelines.

License

Copyright (c) 2025 AudioKeep. All rights reserved.

This is proprietary software. See LICENSE.md for details.

Support

Roadmap

Phase 1 (Q1 2025) - MVP

  • Core audio processing features
  • User authentication and credits system
  • Basic UI with upload/download
  • Stripe payment integration

Phase 2 (Q2 2025) - Enhancement

  • Subscription tiers
  • Batch processing
  • Advanced visualizations
  • Email notifications

Phase 3 (Q3 2025) - Professional

  • Forensic analysis tools
  • API for enterprise
  • Team accounts
  • Priority processing queue

Phase 4 (Q4 2025) - Scale

  • Mobile applications
  • DAW plugin integrations
  • White-label solutions
  • Advanced analytics

Authors

AudioKeep Development Team

Acknowledgments

  • Resemble AI for Resemble Enhance
  • Meta/Facebook Research for Demucs
  • Open-source audio processing community
  • All contributors and testers

Built with ❤️ for audio preservation professionals worldwide

About

AI audio restoration toolkit: noise reduction, declipping, spectral repair, and source separation. Python.

Topics

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages