YouTube Comments Scraper

Paste a list of YouTube videos and get their comments back as a clean table — one row per comment, with the text, who wrote it, how many likes and replies it drew, and when it was posted. Accepts watch URLs, youtu.be share links, Shorts links and bare 11-character video IDs.

No API key. No YouTube Data API quota. No browser automation.

Top comments or newest — your choice

Leave sortByNewest off and you get YouTube's own Top comments ranking: the ones with real engagement behind them, which is what you want for sentiment and research.

Turn it on and you get the most recent comments instead — pulled through the same "Newest first" switch the site itself uses, not a client-side re-sort of the top ones. That is the mode for monitoring a video after launch.

What you get

Field	Example
source_input	https://www.youtube.com/watch?v=dQw4w9WgXcQ
video_id	dQw4w9WgXcQ
comment_id	UgytxzyUgymSYxND1FZ4AaABAg
text	(full comment text)
author_name	@somebody
author_channel_id	UCuAXFkgsw1L7xaCfnd5JJOw
author_is_creator	false
like_count	300000
reply_count	962
published_text	2 weeks ago
published_at	2026-08-03T15:37:42+00:00
author_is_verified	true
author_avatar	https://yt3.ggpht.com/…
reply_level	0
fetched_at	2026-08-15T19:44:02+00:00

like_count and reply_count are real integers, and compact counts are expanded properly — a comment with 300K likes comes back as 300000, not 300.

published_at is an actual timestamp. YouTube only ever shows comments as 2 weeks ago, which you cannot sort, filter or join on. This Actor converts that label to an ISO-8601 UTC timestamp so your rows land in a spreadsheet or database ready to work with. It is approximate to the same granularity YouTube gives — published_text is kept alongside it so you always see the original.

author_is_creator flags the channel owner and author_is_verified flags verified accounts, so you can separate creator responses, brand accounts and ordinary viewers in one filter.

Input

{
  "videoUrls": [
    "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
    "https://youtu.be/ix9cRaBkVe0",
    "8KCuHHeC_M0"
  ],
  "maxCommentsPerVideo": 100,
  "sortByNewest": false,
  "onlyCommentsAfter": "2026-03-01",
  "minLikes": 5
}

Option	Default	Notes
videoUrls	(required)	Any mix of URL formats or bare IDs.
maxCommentsPerVideo	50	1–500 top-level comments per video.
sortByNewest	false	Off = Top comments. On = most recent first.
onlyCommentsAfter	(none)	Date cutoff, e.g. 2026-03-01. Switches sorting to newest-first automatically.
minLikes	0	Drop low-engagement comments.

What people use it for

• Audience research — read what people actually say under the videos in your category, at volume, instead of scrolling one video at a time.
• Product and content feedback — pull comments on your own uploads and sort by like_count to find the objections and requests that resonate.
• Launch monitoring — sortByNewest on a schedule gives you a rolling feed of new reactions.
• Creator response tracking — filter author_is_creator to see how a channel engages.

Export

Results land in a standard Apify dataset: CSV, JSON, Excel or XML download, or straight from the API.

Notes

• Top-level comments only. Reply threads are not expanded — reply_count tells you how many exist.
• Videos with comments disabled come back as rows with ok: false and a plain-English error, tagged with the URL you supplied, so nothing disappears silently.
• Comment text is returned as published, without translation or sentiment scoring.