YouTube Has a Hidden API That Needs No API Key — Here Is How to Use It

YouTube's official Data API has strict quota limits (10,000 units/day). But there's another way.

The Innertube API

YouTube.com itself loads all data through an internal API called Innertube. When you scroll down to load comments, it calls:

POST https://www.youtube.com/youtubei/v1/next

With a JSON body containing the video ID and a continuation token.

How Comments Are Structured (2025+ Format)

YouTube recently changed their comment data format. Comments are now stored in:

frameworkUpdates.entityBatchUpdate.mutations[].payload.commentEntityPayload

Each mutation with type commentEntityPayload contains:

• properties.content.content — comment text
• properties.authorButtonA11y — author name
• toolbar.likeCountNotliked — like count
• properties.publishedTime — relative time
• toolbar.replyCount — number of replies

Getting the Continuation Token

Fetch the video page, extract ytInitialData from the HTML, then navigate:

contents.twoColumnWatchNextResults.results.results.contents
  → itemSectionRenderer.contents
  → continuationItemRenderer.continuationEndpoint.continuationCommand.token

Key Benefits vs Official API

	Official API	Innertube API
API Key	Required	Not needed
Quota	10K units/day	No quota
Rate limit	Yes	Reasonable use
Comment format	Legacy	Latest

I built a YouTube Comments Scraper using this approach — free on Apify Store (search knotless_cadence). Also built YouTube Channel and Search scrapers using the same technique.

Has anyone else worked with the Innertube API?