YouTube Comments Scraper — Comments, Replies & Likes

YouTube Comments Scraper extracts the comments of any public video on youtube.com and returns one flat row per comment: text, like count, number of replies, when it was posted, whether it is pinned, edited or carries the creator's heart, and (optionally) the replies under each comment. Choose YouTube's own Top or Newest order, paste as many videos as you like (watch links, youtu.be, Shorts or live links, or bare IDs) and pay only per comment saved. No YouTube API key, no Google account, no browser.

Privacy by default: commenters are private persons, so their names, channel IDs and channel links are not saved unless you switch hashAuthors off. Instead every commenter gets a stable 16-character pseudonym (authorHash), which is enough to count unique commenters or to see the same person across videos. E-mail addresses and phone numbers inside comment text are removed.

This Actor is part of our YouTube family: YouTube Videos Scraper (search, playlists, single videos with details), YouTube Shorts Scraper and YouTube Channel Scraper. The videos scraper deliberately leaves comments out; this one is built around them.

What data does it return?

Field	Example	Notes
videoId, videoUrl, videoTitle	dQw4w9WgXcQ, …/watch?v=dQw4w9WgXcQ, Rick Astley - Never Gonna Give You Up …	the video the comment belongs to
videoChannelName, videoChannelId	Rick Astley, UCuAXFkgsw1L7xaCfnd5JJOw	the channel that published the video
videoCommentCount	2460059	total comments YouTube reports for the video
commentId, commentUrl	UgzKYksqeQ5gA3t0g0F4AaABAg, …&lc=UgzKYksqeQ5gA3t0g0F4AaABAg	the URL opens the video with that comment highlighted
text	I Rick rolled my entire French class…	full text; e-mails and phone numbers replaced by [email removed] / [phone removed]
likeCount	104000	number; YouTube rounds large counts (104K)
replyCount	621	number of replies in the thread (0 for replies)
publishedText, publishedAt	6 years ago, 2020-10-20T00:00:00.000Z	publishedAt is computed from the relative text, so it is approximate (minute precision for fresh comments, day or coarser for old ones)
isEdited, isPinned, isHearted	true, false, false	edited by the author; pinned by the creator; marked with the creator's heart
isReply, parentCommentId	true, UgzKYksqeQ5gA3t0g0F4AaABAg	replies point at the comment they answer
authorIsChannelOwner, authorIsVerified	false, false	the commenter is the video's channel / has a verification or artist badge
authorHash	5d9765b4cf5af595	stable pseudonym (SHA-256 based), same commenter → same hash in every run
author, authorChannelId, authorUrl	@handle, UC…, https://www.youtube.com/@handle	only when hashAuthors is off; otherwise null
position, sortedBy	1, top	rank in the video's comment list (for replies: rank inside the thread) and the order used
inputUrl, scrapedAt	your input value, ISO time

Use cases

• Audience research and sentiment analysis — pull thousands of comments on a product launch, a trailer or a news clip and run them through your own NLP or an LLM.
• Creator analytics — compare what viewers say under your videos and your competitors' videos; isHearted/isPinned show how creators engage, likeCount shows what resonates.
• Social listening and brand monitoring — schedule a daily run with sort: newest on the videos that matter to you and get only what was posted since.
• Research datasets — pseudonymous comment corpora (hashed authors) for academic or AI work without collecting names.
• Community management — export every comment of a campaign video to CSV or Excel, including replies, to review or answer them.

How to use it

1. Paste one or more Videos (links or IDs).
2. Set Max comments (total) and, if you have several videos, Max comments per video.
3. Pick the Order (Top comments or Newest first) and switch Include replies on if you need them.
4. Keep Hide commenter identities on unless you have a lawful basis to store the commenters' names.
5. Run it and download the dataset as JSON, CSV, Excel or HTML, or read it through the API.

Input example: the prefilled run (20 top comments of one video)

{
  "videoUrls": ["https://www.youtube.com/watch?v=dQw4w9WgXcQ"],
  "maxComments": 20
}

Input example: newest comments with replies from several videos, named authors

{
  "videoUrls": [
    "https://www.youtube.com/watch?v=0e3GPea1Tyg",
    "https://www.youtube.com/shorts/Y8IJi7iP5vk",
    "https://youtu.be/dQw4w9WgXcQ"
  ],
  "sort": "newest",
  "includeReplies": true,
  "maxRepliesPerComment": 3,
  "maxComments": 300,
  "maxCommentsPerVideo": 100,
  "hashAuthors": false
}

Input example: a big pseudonymous dataset from one video

{
  "videoUrls": ["aircAruvnKk"],
  "maxComments": 5000,
  "sort": "top",
  "includeReplies": true,
  "maxRepliesPerComment": 50
}

Output example

A real record from a local test run on 19 October 2026:

{
  "videoId": "dQw4w9WgXcQ",
  "videoUrl": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
  "videoTitle": "Rick Astley - Never Gonna Give You Up (Official Video) (4K Remaster)",
  "videoChannelName": "Rick Astley",
  "videoChannelId": "UCuAXFkgsw1L7xaCfnd5JJOw",
  "videoCommentCount": 2460059,
  "commentId": "UgzKYksqeQ5gA3t0g0F4AaABAg",
  "commentUrl": "https://www.youtube.com/watch?v=dQw4w9WgXcQ&lc=UgzKYksqeQ5gA3t0g0F4AaABAg",
  "text": "I Rick rolled my entire French class because we had to study a city in France and make a quiz about it …",
  "likeCount": 104000,
  "replyCount": 621,
  "publishedText": "6 years ago",
  "publishedAt": "2020-10-20T00:00:00.000Z",
  "isEdited": true,
  "isPinned": false,
  "isHearted": false,
  "isReply": false,
  "parentCommentId": null,
  "authorIsChannelOwner": false,
  "authorIsVerified": false,
  "author": null,
  "authorChannelId": null,
  "authorUrl": null,
  "authorHash": "5d9765b4cf5af595",
  "position": 1,
  "sortedBy": "top",
  "inputUrl": "https://www.youtube.com/watch?v=dQw4w9WgXcQ",
  "scrapedAt": "2026-10-19T19:22:26.089Z"
}

A SUMMARY record in the key-value store lists every video with its status (ok, comments-off, unavailable), how many comments and replies it gave, the proxy used and any errors.

Pricing

Pay per event: each comment or reply saved is charged (comment-result); proxy and platform usage are included.

Duplicates are never saved twice, videos with comments turned off cost nothing, and if you set a maximum cost per run the Actor stops cleanly when it is reached. See the Pricing tab for current prices; paid Apify plans get lower rates.

How it compares

• YouTube Data API v3 (commentThreads): official and exact, but you need a Google Cloud project and an API key, each 100 comments cost quota, the default daily quota is 10,000 units, and it returns commenter names whether you want them or not. This Actor reads the same public comment section a visitor sees, needs no key and hides commenters by default.
• Other comment scrapers in the Store: this one returns the pinned/hearted/edited flags and a stable commenter pseudonym, and lets you cap replies per thread.
• Limits worth knowing: like and reply counts are rounded by YouTube above 1,000 (104K → 104000); dates are relative, so publishedAt is approximate; the comment section is loaded 20 comments per request, so 10,000 comments means about 500 requests.

Integrations

Call it from the Apify API, the JavaScript or Python client, or schedule it in Apify Console. Send the results to n8n, Make, Zapier, Google Sheets, Airtable, Slack or a webhook. AI agents (Claude, ChatGPT, Cursor and other MCP clients) can run it through the Apify MCP server, e.g. "get the 200 newest comments of this video and summarise the sentiment".

FAQ

Is it legal to scrape YouTube comments? The Actor only reads comments that YouTube shows to logged-out visitors. Comments are user-generated content and the commenters are usually private persons, so data-protection laws (GDPR, CCPA, Chile's Ley 21.719 and others) can apply to how you store and use the results. That is why names and channel IDs are hidden by default and contact details are stripped from the text. Only turn hashAuthors off when you have a lawful basis. This is not legal advice.

What exactly is authorHash? The first 16 hexadecimal characters of a SHA-256 digest of the commenter's channel ID (with a fixed prefix). It is the same in every run, so you can count unique commenters or join datasets, but it cannot be turned back into a name.

Do I need a proxy? No setup needed. The Apify datacenter proxy is used by default; if YouTube blocks it twice in a row, the run switches to Apify residential proxy (country = region) for the rest of the run.

Why does a video return 0 comments? Comments are turned off on that video (the SUMMARY record says comments-off), the video is private or removed (unavailable), or the link was not a video at all (channel and playlist links are skipped with a warning).

Can I get only the comments posted since my last run? Use sort: "newest" with a reasonable maxComments, then filter on publishedAt or on the commentIds you already have.

Does it download videos or avatars? No. Only comment data; avatar URLs are not saved.

Which languages work? Comments come back in whatever language they were written. Keep language at en so that relative dates and rounded counts are parsed reliably.

Changelog

• 2026-10-19 — First version: comments and replies of any video, Top / Newest order, per-video and per-thread caps, pinned / hearted / edited flags, hashed commenters by default, contact scrubbing, pay per comment.

En español

YouTube Comments Scraper extrae los comentarios públicos de cualquier video de YouTube sin API key ni login: texto, likes, cantidad de respuestas, fecha aproximada, si está fijado, editado o tiene el corazón del creador, y opcionalmente las respuestas de cada hilo, en orden "Principales" o "Más recientes". Por defecto no guarda el nombre ni el canal de quien comenta: cada persona recibe un seudónimo estable (authorHash) y se eliminan correos y teléfonos del texto, pensando en la Ley 21.719 y el RGPD. Se cobra por cada comentario guardado, con proxy incluido. Consulta la pestaña Pricing para ver los precios actuales; los planes de pago de Apify tienen tarifas más bajas. Para videos, listas y canales usa nuestros YouTube Videos Scraper, YouTube Shorts Scraper y YouTube Channel Scraper.