How to Scrape YouTube Data Without the Official API (No Quota Limits)

#youtube #api #webdev #javascript

The YouTube Data API has strict quota limits — 10,000 units per day, and a single search costs 100 units. That gives you only 100 searches per day.

Here is how to get YouTube data without those limits, using the internal Innertube API and direct page parsing.

4 YouTube Data Tools

1. YouTube Comments Scraper

The official API charges 1 unit per comment. With 10,000 daily units, you can only get 10,000 comments per day.

The Innertube API has no such limits. Here is the approach:

// Fetch the video page
const html = await fetch(`https://www.youtube.com/watch?v=${videoId}`).then(r => r.text());

// Extract the continuation token from ytInitialData
const ytData = JSON.parse(html.match(/ytInitialData\s*=\s*({.*?});/)[1]);

// Navigate to comments section
const contents = ytData.contents.twoColumnWatchNextResults
  .results.results.contents;

// Find the comment section continuation
const commentSection = contents.find(c => c.itemSectionRenderer?.targetId === "comments-section");

The scraper handles pagination automatically, extracting all comments with:

Author name and channel
Comment text
Like count
Reply count
Published date
Whether the author is verified

2. YouTube Channel Scraper

Get complete channel data without API:

Subscriber count
Total videos
Channel description and links
Recent video list with view counts
Channel creation date

3. YouTube Search Scraper

Search YouTube and get structured results:

Video title, URL, thumbnail
View count and publish date
Channel name
Duration
Description snippet

4. YouTube Transcript Scraper

Extract video captions and subtitles:

Supports multiple languages
Includes timestamps
Works with auto-generated captions
Handles YouTube Shorts

// Get caption tracks from player response
const captionMatch = html.match(/"captionTracks":\[(.+?)\]/);
const tracks = JSON.parse(`[${captionMatch[1]}]`);

// Fetch the transcript XML
const transcript = await fetch(track.baseUrl).then(r => r.text());