SoundCloud Scraper for AI Training

A Node.js-based SoundCloud scraper designed for collecting audio data to train AI models. Includes three specialized scripts for music generation, audio enhancement, and speech processing use cases.

Features

• Playlist scraping. Extract metadata from SoundCloud playlists including artist names, track titles, play counts, and URLs.
• Search results scraping. Find Creative Commons licensed tracks with download availability filtering.
• Profile scraping. Collect podcast episodes and long-form spoken content from user profiles.
• Proxy support. Built-in residential proxy integration for reliable, undetected scraping.
• Headless browser automation. Uses Puppeteer to handle JavaScript-heavy pages and dynamic content.
• Download detection. Automatically identifies tracks with enabled download buttons.

Installation

Prerequisites

• Node.js 14.0+
• Decodo residential proxies

Setup

git clone https://github.com/Decodo/soundcloud-scraper.git
cd soundcloud-scraper
npm install

Configure proxies

Get your proxy credentials from the Decodo dashboard and update them in each script file:

await page.authenticate({
  username: 'YOUR_PROXY_USERNAME',  // Replace with your username
  password: 'YOUR_PROXY_PASSWORD'   // Replace with your password
});

Usage

1. Music generation AI training

Train models like Suno AI, AIVA, or Stable Audio using curated playlists that represent successful musical patterns across different genres.

File: music-generation.js

Scrape trending playlists to collect metadata for training music generation models.

node music-generation.js

What it does:

• Targets SoundCloud playlist pages
• Extracts artist names, track titles, play counts
• Outputs structured data with rankings

Customize the target:

// Edit line 43 in music-generation.js
await page.goto('https://soundcloud.com/YOUR-PLAYLIST-URL', {

Output example:

Found 50 playlist tracks

1. Artist Name - "Track Title" (1.2M plays)
   https://soundcloud.com/artist/track

2. Audio enhancement AI training

Collect Creative Commons tracks to train models that clean degraded recordings, remove noise, and restore audio quality like Adobe Enhance Speech or Descript.

File: audio-enhancement.js

Find Creative Commons tracks with download availability for audio cleanup model training.

node audio-enhancement.js

What it does:

• Searches SoundCloud with custom queries
• Auto-scrolls to load more results
• Filters tracks with download buttons enabled

Customize the search:

// Edit line 49 in audio-enhancement.js
await page.goto('https://soundcloud.com/search/sounds?q=YOUR-SEARCH-QUERY', {

Output example:

Total items found: 85
Found 42 downloadable tracks:

1. Artist Name - "Track Title"
   https://soundcloud.com/artist/track

3. Speech/voice AI training

Extract podcast episodes to train speech recognition, voice cloning, and natural language processing models like Whisper (OpenAI) or ElevenLabs.

File: voice-training.js

Extract podcast episodes and lectures for speech recognition and voice AI training.

node voice-training.js

What it does:

• Scrapes user profile pages
• Focuses on long-form spoken content
• Identifies downloadable episodes

Customize the target:

// Edit line 43 in voice-training.js
await page.goto('https://soundcloud.com/YOUR-PROFILE/tracks', {

Configure limits:

// Edit lines 62-63 in speech-training.js
const maxTracksToScrape = 50;  // Maximum tracks to process
const maxScrollAttempts = 20;  // Scroll depth

Output example:

Found 68 total tracks
24 downloadable tracks:

1. Podcast Name - "Episode Title"
   https://soundcloud.com/podcast/episode

Configuration

Each script includes configurable parameters at the top of the file:

Proxy settings:

'--proxy-server=http://gate.decodo.com:7000'  // Proxy endpoint
username: 'YOUR_PROXY_USERNAME'               // Your credentials
password: 'YOUR_PROXY_PASSWORD'

Scraping behavior:

headless: true          // Run browser invisibly
timeout: 45000          // Page load timeout (ms)
maxScrollAttempts: 5    // Pagination depth
targetResults: 500      // Result limit

Resource blocking:

// Scripts automatically block images, media, fonts
// to improve performance and reduce bandwidth

Best practices

• Start with small limits (10-20 items) to test your setup
• Use residential proxies to avoid IP bans
• Add delays between requests to respect rate limits
• Monitor console output for errors and warnings
• Store credentials securely, never commit them to Git

Proxy setup

Get residential proxies from Decodo:

1. Sign up at dashboard.decodo.com
2. Navigate to Residential Proxies → Proxy Setup
3. Copy your Username and Password
4. Update credentials in each script file

Troubleshooting

Browser won't launch?

• Verify Node.js 14+ is installed
• Install Puppeteer: npm install
• Try setting headless: false to see browser

Getting blocked?

• Check proxy credentials are correct
• Increase delays between requests
• Verify proxy quota on Decodo dashboard

No results found?

• SoundCloud's HTML may have changed
• Check target URL is accessible
• Review browser console output

Documentation

• Blog post tutorial
• Decodo proxy setup
• Puppeteer documentation

Related projects

🗺️ Google Maps Scraper

🔍 Google Lens Scraper

📰 Google News Scraper

💬 Reddit Scraper