System Overview

Understanding the dual-system architecture of v9.3

CopyrightChains v9.3 (CC93) Architecture

v9.3 consists of TWO COMPLEMENTARY SYSTEMS working together to provide comprehensive music verification and fraud detection.

System 1: AI Detection

Identifies AI-generated music from 8+ platforms

System 2: Playlist Intelligence

Detects streaming fraud and bot activity

System 1: AI Music Generation Detection

Detects whether tracks were created by AI platforms (Suno, Udio, AIVA, etc.) or human artists.

Supported AI Platforms

Suno

Risk: 85% copyright, 70% legal, 90% ownership

Udio

Risk: 80% copyright, 65% legal, 85% ownership

AIVA

Risk: 40% copyright, 30% legal, 45% ownership

Soundraw

Risk: 35% copyright, 25% legal, 40% ownership

Soundful

Risk: 35% copyright, 25% legal, 40% ownership

Mubert

Risk: 45% copyright, 35% legal, 50% ownership

Beatoven

Risk: 30% copyright, 25% legal, 35% ownership

Endel

Risk: 25% copyright, 20% legal, 30% ownership

Technology Stack

  • Multi-AI Analysis: OpenAI GPT-4, Anthropic Claude Opus 4.1, Perplexity Sonar
  • Audio Processing: librosa (Python audio analysis library)
  • Metadata APIs: Spotify Web API, YouTube Data API v3, Deezer API
  • Real-time Communication: WebSocket (port 8766)
  • Database: MariaDB via MCP toolbox

Analysis Process

  1. Track Selection: User selects track from database
  2. Metadata Enrichment: Fetch data from Spotify, YouTube, Deezer
  3. Audio Download: Retrieve 30-second preview (if available)
  4. Feature Extraction: Analyze spectral, temporal, and timbral features
  5. Multi-AI Analysis: Query 3 AI platforms for platform identification
  6. Risk Assessment: Calculate confidence-scaled risk scores
  7. Results Display: Show platform likelihood and risk breakdown

System 2: Playlist Intelligence & Fraud Detection

Detects streaming fraud, bot activity, and fake engagement in Spotify playlists.

Technology Stack

  • Machine Learning: XGBoost (fraud detection), Ensemble (artist authenticity)
  • Statistical Analysis: Benford's Law, time-series anomalies, clustering
  • Behavioral Analytics: Bot pattern detection, coordinated activity
  • Network Analysis: Graph-based fraud network mapping
  • Engagement Metrics: 50+ metrics across 7 dimensions

Fraud Detection Dimensions

1. Stream Velocity

Sudden spikes vs organic growth patterns

2. Engagement Quality

Skip rates, completion rates, listener diversity

3. Curator Authenticity

5-tier classification system

4. Network Analysis

Coordinated bot networks and farms

5. Behavioral Patterns

Bot listening signatures vs human variance

6. Artist Authenticity

Cross-platform verification and history

7. Financial Impact

Estimated revenue from fraudulent streams

Playlist Fraud Score (0-100)

  • 0-20: Clean (organic engagement)
  • 21-40: Low risk (minor irregularities)
  • 41-60: Moderate risk (investigation recommended)
  • 61-80: High risk (likely fraud)
  • 81-100: Critical risk (definite fraud)

Unified Track + Playlist Verification

When both systems are complete, you can analyze tracks and playlists together for comprehensive verification.

Combined Workflow

  1. Track Selection: Choose track from database
  2. AI Detection: System 1 analyzes generation likelihood
  3. Playlist Analysis: System 2 analyzes playlists containing track
  4. Network Mapping: Identify coordinated fraud involving AI tracks
  5. Combined Risk Score: Aggregate risk assessment
  6. Evidence Package: Generate court-admissible report

Use Case 1

Independent artist verifies authenticity before signing with label

Use Case 2

Rights holder detects AI content theft and files dispute

Use Case 3

Playlist curator ensures legitimate engagement metrics

Use Case 4

Legal team compiles evidence for copyright infringement case

Technical Architecture

Services

  • enhanced_ai_detector.py: Main verification service
  • audio_analyzer.py: librosa feature extraction
  • api_enrichment.py: Spotify/YouTube/Deezer integration
  • playlist_processor.py: Fraud detection engine (planned)
  • fraud_detector.py: ML model inference (planned)

Infrastructure

  • MCP Server: Port 8091 (WebSocket), 5-second polling
  • Processor Service: Port 8766 (WebSocket), Python 3.11
  • Web UI: Port 80 (HTTP), Vanilla JavaScript
  • Database: MariaDB external connection

Docker Containers

  • cc93-mcp-server: Node.js + MCP toolbox
  • cc93-processor: Python 3.11 + librosa
  • cc93-webui: Nginx static server

v9.2 vs v9.3 Comparison

v9.2 (Production) Focus:

  • Blockchain-based copyright registration
  • Musical works and recordings management
  • 16/84 royalty distribution
  • Wyoming Series LLC creation
  • Investigation and dispute resolution

v9.3 (Beta) Focus:

  • AI music generation detection
  • Playlist intelligence and fraud detection
  • Real-time verification
  • Machine learning fraud detection
  • Network analysis for coordinated fraud

Integration

v9.3 complements v9.2 by adding:

  • Pre-registration verification: Detect AI before registering
  • Ongoing monitoring: Detect fraud in registered works
  • Enhanced dispute evidence: AI generation proof
  • Rights protection: Identify infringement of AI-generated content

Related Topics