jesusyanez / free-instrumental-api

An API for youtube video metadata.

Key Features

1. YouTube metadata scraping (w/o YouTube API)
2. Next.js API w/ Serverless Functions

Stack

• JavaScript
• Node
• Express
• Next.js

Scraper

• Scrapes youtube search for metadata and creates a JSON file.
• Made an artists array and looped the script to search for each array item, so I would get one final array.
• Quality Control: Filtered to only save data for videos over 30k views.
• Variety: Right now the data includes metadata of 850+ YouTube videos (30+ artists).

API

• Has one endpoint which returns a random object.
• Used it for my free rap beat app
• At first I made it with TypeScript and Express.
• Changed it to a Next.js API to utilize Serverless Functions on Vercel.com

Usage

HTTPS

GET https://free-instrumental-api.vercel.app/api/beats

JSON Response

{
  "0": {
    "description": "",
    "duration": 132000,
    "duration_raw": "2:12",
    "uploaded": "11 months ago",
    "views": 1044475,
    "channel": {
      "name": "ODDstatus",
      "link": "https://www.youtube.com/channel/UCgIXpunYoqLMyZiPjyPwr6A",
      "verified": false
    },
    "id": "GSb3JGYSEI4",
    "link": "https://www.youtube.com/watch?v=GSb3JGYSEI4",
    "thumbnail": "https://i.ytimg.com/vi/GSb3JGYSEI4/maxresdefault.jpg",
    "title": "[FREE] JID Type Beat \"The Recreation\"",
    "shareLink": "https://youtu.be/GSb3JGYSEI4"
  }
}

To-Do

• Scrape Youtube search results
  • Use requests
  • No Google/YouTube API
• Create Rest API that sends random video metadata object
• Automate scraper with cronjob maybe
• Add custom queries:
  • Genres: Trap, Hyperpop, Dark
  • Artists
• Clean up scraped data to reduce API bandwidth & response times