CoolFace
Apppublic

hardbanrecords/Metadata-Engine

sourceHugging Faceotherupdated 8mo agoView on Hugging Face
0likes
App README

๐ŸŽต Music Metadata Engine: Professional Edition

AI-Powered Audio Analysis & Metadata Enrichment Platform

A comprehensive music metadata extraction and enrichment system that combines DSP audio analysis (Essentia/Librosa) with AI-powered classification (Gemini 2.0), delivering professional-grade metadata for music production, distribution, and licensing.


โœจ Features

๐Ÿ”ฌ Core Analysis

  • โ€”BPM Detection: Precise tempo analysis using Essentia RhythmExtractor
  • โ€”Key Detection: Chromatic key & mode detection (12-tone + major/minor)
  • โ€”Audio Metrics: LUFS loudness, danceability, energy level
  • โ€”SHA-256 Fingerprinting: File integrity verification

๐Ÿค– AI Enrichment (Gemini 2.0 Flash)

  • โ€”Genre Classification: Main + additional genres with confidence scores
  • โ€”Mood & Vibe: Context-aware emotional classification
  • โ€”Instrumentation Detection: AI-identified instruments from audio signature
  • โ€”Use Case Tagging: Commercial sync licensing categories
  • โ€”Vocal Style Analysis: Gender, timbre, delivery, emotional tone

๐Ÿ“ฆ Export Formats

  • โ€”CSV (MP3Tag): Bulk import compatible with MP3Tag software
  • โ€”JSON: Full metadata structure with ISO timestamps
  • โ€”DDEX ERN 4.3: XML for professional music distribution (Spotify, Apple Music)
  • โ€”CWR V2.1: Common Works Registration for PRO (ASCAP, BMI, GEMA)

๐ŸŒ IPFS Certificate

  • โ€”JSON-based certificate uploaded to Pinata (IPFS)
  • โ€”PDF-style viewer with legal disclaimer & verification URL
  • โ€”Cryptographic proof of file existence (SHA-256 hash)
  • โ€”Berne Convention compliant for "Prior Art" claims

๐Ÿš€ Advanced Features

  • โ€”Batch Processing: Concurrent analysis of multiple files (3x faster)
  • โ€”AI Cover Art Generator: Pollinations.ai with gradient fallback
  • โ€”Premium UI/UX: Glassmorphic design with dark mode

๐ŸŽฏ Use Cases

  • โ€”Music Producers: Accurate BPM/key for mixing & DJing
  • โ€”Labels & Distributors: DDEX-compliant metadata for global distribution
  • โ€”Sync Licensing: Mood/use case tagging for music supervisors
  • โ€”Copyright Protection: IPFS certificates for proof of existence
  • โ€”PRO Registration: CWR exports for royalty collection

๐Ÿ› ๏ธ Technology Stack

Backend:

  • โ€”FastAPI (Python 3.10)
  • โ€”Essentia (DSP audio analysis)
  • โ€”Librosa (fallback DSP)
  • โ€”Gemini 2.0 Flash Experimental (AI enrichment)
  • โ€”Mutagen (metadata tagging)

Frontend:

  • โ€”React 19.2
  • โ€”TypeScript 5.2
  • โ€”Vite 5.2
  • โ€”Tailwind CSS 3.4

Infrastructure:

  • โ€”Hugging Face Spaces (Docker)
  • โ€”Pinata (IPFS storage)
  • โ€”SQLite (job queue)

๐Ÿš€ Quick Start

  1. 1.Upload an audio file (.mp3, .wav, .flac)
  2. 2.Analyze - wait 30-45 seconds for DSP + AI processing
  3. 3.Review metadata across 5 result cards
  4. 4.Export in your preferred format (CSV/JSON/DDEX/CWR)
  5. 5.Protect your work with IPFS certificate

๐Ÿ“Š API Endpoints

Analysis

  • โ€”POST /analyze - Full track analysis (DSP + AI)
  • โ€”POST /mir/analyze - MIR-only analysis
  • โ€”POST /mir/batch_analyze - Batch processing

Export

  • โ€”GET /export/csv/{job_id} - MP3Tag CSV
  • โ€”GET /export/json/{job_id} - Full JSON
  • โ€”GET /export/ddex/{job_id} - DDEX ERN 4.3 XML
  • โ€”GET /export/cwr/{job_id} - CWR V2.1

Generative

  • โ€”POST /generate/cover - AI cover art
  • โ€”POST /generate/hash - SHA-256 hash
  • โ€”POST /generate/certificate - IPFS certificate

Tagging

  • โ€”POST /tag/mp3 - ID3 tagging
  • โ€”POST /tag/flac - FLAC tagging

Full API documentation: /docs (FastAPI Swagger)


๐Ÿ” Environment Variables

Required for full functionality:

bash
GEMINI_API_KEY=your_gemini_key
PINATA_JWT=your_pinata_jwt
PINATA_GATEWAY=https://gateway.pinata.cloud/ipfs/

Optional (enrichment):

bash
GROQ_API_KEY=your_groq_key  # AI fallback
SPOTIFY_CLIENT_ID=your_spotify_id
LASTFM_API_KEY=your_lastfm_key

๐Ÿ“ˆ Performance

  • โ€”Analysis Speed: 30-45 seconds per track (3-5 min duration)
  • โ€”Batch Processing: 3x faster with concurrent queue (3 parallel jobs)
  • โ€”Supported Formats: MP3, WAV, FLAC (up to 100MB)

๐Ÿ“„ License

Proprietary License - ยฉ 2026 HardbanRecords Lab. All Rights Reserved. Commercial use without prior authorization is strictly prohibited.


๐Ÿ™ Acknowledgments

Built with:


๐Ÿ“ž Support

For issues, feature requests, or questions:

  • โ€”Open an issue on GitHub
  • โ€”Contact: hardbanrecords@proton.me

Version: 2.1.0 Status: Production Ready โœ… Last Updated: January 2026