YouTube comment
scraper
Paste a video link and read what its audience says: top comments with likes, replies and dates, the questions they ask and the words they repeat. 5 free comment scrapes a day, no account.
Export YouTube comments without a quota budget
The comments under a video are unprompted audience research. Viewers say what confused them, what they want next, which product they bought instead and what they would pay for. Brands read them on sponsored videos, marketers mine them for ad copy and content ideas, and researchers collect them as a dataset.
This scraper reads up to about 40 top-level comments from any public video and shows 20, with the author, like count, reply count and date. It also counts what you would otherwise tally by hand: the share of comments that ask a question, the share that point to a timestamp, the most-liked comment and the words people repeat. There is no sentiment score, because a plain count is something you can check.
Developers usually hit the same wall: the official YouTube Data API needs a Google Cloud project and spends quota units on every comment page, out of a daily budget shared with the rest of the app. If you want to know whether your comment job fits, run it through the YouTube API quota calculator. The Scavio YouTube API returns the same comments with one API key, 1 credit per page of about 20, with a cursor to page through the whole thread.
How to scrape YouTube comments
- 1
Paste the video
Copy the link from the address bar or the Share button. Watch links, youtu.be short links, Shorts links and bare 11-character video IDs all work.
- 2
Filter and sort if you want
Type a keyword to see only comments that mention it, such as a product name, a competitor or a word like price. Sort by most liked, most replies, newest, or keep YouTube's own Top order.
- 3
Read the comments and the numbers
You get up to 20 comments with author, likes, replies and date, plus how many comments were read, the most-liked comment, the share that ask a question or cite a timestamp, and the words people repeat most.
Frequently asked questions
What does the YouTube comment scraper show?+
Up to 20 comments from the video with the author, a link to their channel, like count, reply count, how long ago it was posted, and whether it was edited or written by the creator. Above the list you see how many comments were read, the most-liked comment, the share of comments that ask a question, the share that cite a timestamp, and the most repeated words.
How many comments does it read?+
Each run reads up to two pages, about 40 top-level comments, computes the stats over all of them and shows 20. The Scavio API returns the same comments about 20 per page with a cursor for the next page, so you can page through every comment on a video from your own code.
Is it free?+
Yes. You get 5 free comment scrapes a day with no account. For more, get an API key: 50 free credits on signup, no card required, and each page of about 20 comments costs 1 credit.
Can I export YouTube comments to CSV or Excel?+
Not from this page. The full comment list is available via the API as JSON, which you can write to a CSV, a Google Sheet or a database with a few lines of code. Each comment carries comment_id, text, like_count, reply_count, published_time and an author object.
Can I sort by newest comments?+
The tool can re-sort the comments it reads by newest, most liked or most replies. The pages themselves arrive in YouTube's default Top comments order, so on a video with thousands of comments, newest here means the newest of the comments read, not every comment ever posted.
How is the question share calculated?+
A comment counts as a question when it contains a question mark or opens with a word like how, what, why, where, when, who, which, can, does, is or should. It is a plain text rule, not an AI model, so you can check every tagged comment yourself.
Does it do sentiment analysis?+
No. The tool shows word counts, questions and timestamps, which you can verify by reading the comments. If you want sentiment, pull the comments through the API and run them through the model you already trust.
Does it get replies to comments?+
The tool shows how many replies each comment has. Through the API, each comment includes a reply_cursor, and POST /api/v1/youtube/comments/replies with that cursor returns the replies, 1 credit per page.
Why not use the YouTube Data API?+
You can. As of October 2026, Google's commentThreads.list costs 1 quota unit per call and returns up to 100 comment threads, against a default of 10,000 units a day shared with your other calls, and you need a Google account, a Cloud project and credentials first. The Scavio API needs one API key, has no daily unit budget to plan around, and uses the same key for Shorts, transcripts, channels and other platforms.
Can I use this in my own code?+
Yes. POST /api/v1/youtube/comments with {"video_id": "dQw4w9WgXcQ"} returns data.comments, data.next_cursor and data.has_more. Send the next_cursor back as cursor to get the next page. Each call costs 1 credit.
Need what was said in the video itself? Try the YouTube transcript downloader, or list a channel's uploads with the YouTube channel scraper.
Pull every comment with the YouTube API
This tool is one endpoint on the Scavio API. Page through every comment on a video, fetch replies, and pair them with transcripts and video stats in your own pipelines and agents.