Youtube Comments Extractor

Extract every comment from a YouTube video, including full reply threads, without touching the YouTube Data API or dealing with its quota limits. Give the actor a video URL and it returns comment text, author details, like counts, reply counts, and publish dates as clean, structured data ready for a spreadsheet, a dashboard, or your own pipeline.

Why use this actor

The YouTube Data API caps free usage at a small daily quota, requires an API key and OAuth setup, and still won't hand you a video's comments in one clean call. This actor skips all of that. Paste a video URL, choose how many comments you want, and it walks YouTube's own comment feed the same way a browser would, following pagination automatically until it hits your limit or runs out of comments.

No YouTube API key, no OAuth consent screen, no quota to track. Just a video URL and an output you can open in Sheets five minutes from now.

Who this is for

• SEO and content teams mining comment threads for the questions, phrases, and pain points an audience is already typing out, then turning that into content briefs and FAQ sections
• Marketers and brand teams tracking what people are saying about a product, campaign, or competitor on any public video, including videos they don't own
• Data analysts and researchers feeding comment text into a sentiment or topic model to measure how an audience actually reacted, not just how many likes a video got
• Creators and channel managers reviewing comments and reply threads across a catalog of videos in one batch instead of scrolling through YouTube by hand
• Moderation and compliance teams that need a timestamped, exportable record of comment activity on a video

How it works

1. You provide one or more YouTube video URLs.
2. The actor loads each video's comment section and reads YouTube's own comment feed, sorted by top or newest, exactly as configured.
3. For every top-level comment it collects, it follows the reply thread underneath that comment (when reply collection is on) before moving to the next one, so replies stay grouped with their parent instead of arriving as a separate, disconnected batch at the end.
4. It keeps paging, comment by comment and reply thread by reply thread, until it reaches your comment limit or the video runs out of comments, whichever comes first.

Videos with comments turned off, or comments restricted to signed-in viewers, return zero rows for that URL instead of failing the whole run, so a bad link in a batch never blocks the rest of your list.

Input

Field	Type	Description
startUrls	array of strings	One or more YouTube video URLs. Accepts full watch links, youtu.be short links, and Shorts links
sortBy	string	top or newest. Controls the order comments are fetched in. Default top
includeReplies	boolean	When on, fetches the full reply thread under each comment before moving to the next one. When off, only top-level comments are collected. Default true
maxComments	integer	Maximum number of comments to collect per video, top-level comments and replies combined. Default 100, up to 5000
requestTimeoutSecs	integer	Per-request timeout in seconds. Default 60

Example input:

{
  "startUrls": ["https://www.youtube.com/watch?v=xuCn8ux2gbs"],
  "sortBy": "top",
  "includeReplies": true,
  "maxComments": 200
}

Output

Each dataset row is one comment or one reply. A top-level comment looks like this:

{
  "videoId": "xuCn8ux2gbs",
  "videoUrl": "https://www.youtube.com/watch?v=xuCn8ux2gbs",
  "videoTitle": "history of the entire world, i guess",
  "commentId": "UgzbUWbHj_VFK6pb2_14AaABAg",
  "commentText": "Truly one of the greatest videos of all time",
  "authorName": "@jatootjaboot1578",
  "authorChannelUrl": "https://www.youtube.com/@jatootjaboot1578",
  "authorAvatarUrl": "https://yt3.ggpht.com/...",
  "authorIsVerified": false,
  "authorIsCreator": false,
  "likeCount": 5100,
  "replyCount": 46,
  "publishedTime": "2 months ago",
  "isReply": false,
  "position": 1,
  "scrapedAt": "2026-08-16T10:18:57.008045+00:00",
  "error": null
}

A reply carries the same fields plus parentCommentId, pointing back at the comment it replies to:

{
  "commentId": "UgzbUWbHj_VFK6pb2_14AaABAg.AXwxAF9KBLcAXwxYmrDS4T",
  "commentText": "My autism is so obsessed with this video and the various sound bites that I have to return multiple times a year.",
  "isReply": true,
  "parentCommentId": "UgzbUWbHj_VFK6pb2_14AaABAg"
}

Field reference

Field	Description
videoId, videoUrl, videoTitle	Identifies which video the row belongs to
commentId	YouTube's unique ID for this comment or reply
commentText	The comment's full text
authorName, authorChannelUrl, authorAvatarUrl	Who posted it and where to find their channel
authorIsVerified, authorIsCreator	Whether the author has a verified badge, or is the video's own creator replying in their own comments
likeCount	Number of likes on the comment
replyCount	YouTube's own reply count, shown on the top-level comment
publishedTime	YouTube's relative timestamp, e.g. "2 months ago"
isReply	true for a reply, false for a top-level comment
parentCommentId	Only present on replies, pointing to the parent comment's commentId. Top-level comments don't carry this field at all
position	The row's order within the run
scrapedAt	UTC timestamp of when the row was collected
error	null on a normal row. Populated only when a specific URL failed to process

Export any run as JSON, CSV, or Excel, or push it straight to Google Sheets from the Apify dataset view.

Tips

• Add several URLs to startUrls to pull comments from multiple videos in a single run.
• Set sortBy to newest when you're monitoring a video for fresh activity rather than pulling its full history.
• maxComments counts top-level comments and replies together. A video with long reply threads can use up the whole budget on just a handful of top-level comments and their replies, so raise the limit if you need broader top-level coverage across more comments.
• Turn includeReplies off if you only care about the top-level conversation and want to cover more distinct comments per run.

FAQ

Does this work on Shorts? Yes. Any public video with comments enabled works, Shorts included.

Can I scrape multiple videos in one run? Yes. Add as many URLs as you need to startUrls and they're processed one after another in the same run.

Do I need a YouTube account or API key? No. Just a public video URL.

What happens if a video has comments disabled? The actor logs it and moves on to the next URL in your list instead of failing the whole run.

Why does a top-level comment sometimes have fewer replies than replyCount shows? The actor stops adding rows once your run hits maxComments. If a comment thread's replies would push the run past that limit, only the replies that fit are included.