YouTube Comments Scraper

Extract comments from any YouTube video — with the author, like count, reply count, and when each one was posted. Sort by top comments or newest first. Pagination is handled for you: ask for 2,000 comments and you get 2,000.

Why this one returns data. YouTube blocks the datacenter address ranges that scraping platforms run on, which is why so many comment scrapers return empty results or bot-check errors. This one routes every request through a residential connection, so it sees what a normal viewer sees.

Quick start

{
  "videoUrls": ["https://www.youtube.com/watch?v=dQw4w9WgXcQ"],
  "maxCommentsPerVideo": 500,
  "sortBy": "top"
}

What you get per comment

Field	Description
text	The comment itself
author · authorChannelId · authorAvatar	Who wrote it
authorIsCreator	Whether the video's own channel wrote it
likeCountText · likeCount	Likes, as displayed (275K) and parsed where possible
replyCount	Number of replies
publishedText	When it was posted, as YouTube shows it
commentId	Stable identifier
videoId · videoTitle · channel	Which video it came from

By default each comment is its own dataset row, so the output exports straight to CSV or Excel. Turn off One row per comment to get one row per video with the comments nested inside instead.

What people use it for

• Audience research — what viewers actually say about a product, topic, or competitor
• Sentiment analysis — feed comments into a classifier or an LLM
• Content ideas — the questions asked repeatedly under a popular video are a content plan
• Community management — find unanswered questions on your own videos
• Market research — read complaints and requests in a niche, at scale

Pairs well with the YouTube Data Scraper — use it to find videos in a niche, then this one to read what people said about them.

Pricing

Pay per comment returned. A video with 20 comments costs a twentieth of one with 400. Videos that fail or have comments disabled cost you nothing — they still appear in the dataset with an error field so nothing disappears silently.

Notes and limits

• Videos with comments disabled return a clear error rather than an empty result.
• Replies to comments are counted (replyCount) but not expanded into separate rows.
• likeCount is parsed from YouTube's abbreviated display where possible; likeCountText always carries the exact string YouTube shows (275K, 1.2M).
• Each video is fetched independently — one failure never aborts the run.
• This actor reads publicly available data only.

Support

Something not returning what you expect? Open an issue on the Issues tab with the video URL you used.