TikTok Comments Scraper: Comments and Replies from Videos

TikTok Comments Scraper: comments, replies, likes and commenters

Use this to export comments and replies from public TikTok videos for sentiment and audience research.

Price per 1,000 comments	Free plan	Starter	Scale	Business
This Actor	$0.87	$0.70	$0.52	$0.15
Most-used alternative	$1.25	$1.00	$0.75	$0.50

Turn any public TikTok video into a clean table of its comments: comment text, likes, number of replies, date, commenter username, display name, user ID, avatar, pinned and liked-by-creator flags, and optionally the replies under each comment, threaded by parent ID. Paste video URLs. Plain HTTP, no browser, no login, no cookies.

• In short: TikTok Comments Scraper (Apify actor transparent_meteorite/tiktok-comments-scraper) reads the public comments and replies of TikTok videos and returns one row per comment.
• Who it is for: brand and community managers, social listening and sentiment analysis, agencies reporting on campaigns, creators finding their most engaged fans, researchers.
• Input: video URLs (short links and bare IDs too); comments per video; replies per comment; skip pinned; date and minimum-likes filters; max rows; monitor mode; proxy.
• Output: 20+ fields per comment: cid, text, diggCount, replyCommentTotal, createTimeISO, uniqueId, nickname, uid, avatarThumbnail, repliesToId, isPinned, authorLiked, videoWebUrl, commentUrl. JSON, CSV or Excel.
• Pricing: pay per event. Per comment delivered: $0.00087 free plan, $0.0007 Bronze, $0.00052 Silver, $0.00015 Gold, Platinum and Diamond.
• Limits: only what TikTok shows logged-out visitors; comments the creator hid or TikTok filtered are not returned.

Paste a video URL, press Start, get its comments as a table in about 5 seconds.

Run in 30 seconds

1. Open the actor in Apify Console (a free Apify account is enough).
2. Press Start with the prefilled input: 50 comments of a TikTok video, about $0.06 on the free plan.
3. Open the Output tab: the Comments view is a table. Download as CSV, Excel or JSON.

{
  "postURLs": ["https://www.tiktok.com/@tiktok/video/7106594312292453675"],
  "commentsPerPost": 50,
  "maxRepliesPerComment": 0
}

An empty input {} also works: it reads 100 comments of an example video.

Why choose this over alternatives

	This actor	Typical TikTok comment scrapers
Price per comment (free plan)	$0.00087	$0.00125
Price per comment (Gold, Platinum, Diamond)	$0.00015	$0.0005 / $0.0003 / $0.00015
Replies, threaded by parent ID	Yes	Yes
Pinned and liked-by-creator flags	Yes	Partly
Skip pinned, date and minimum-likes filters	Yes, combined	Rarely
Monitor mode (only new comments)	Yes	No
Link straight to each comment	Yes (commentUrl)	No
Videos that cannot be read, filtered comments	Not charged	Varies
Memory	256 MB, no browser	Often a browser

What you get: output fields

Field	Meaning
cid	Comment ID
text	Comment text
diggCount	Likes on the comment
replyCommentTotal	Number of replies (top-level comments)
createTime, createTimeISO	When it was posted
uniqueId, nickname, uid, secUid	Commenter username, display name and IDs
avatarThumbnail, profileUrl	Commenter avatar and profile link
isReply, repliesToId, replyToReplyId	Reply threading
isPinned, authorLiked	Pinned by the creator, liked by the creator
commentLanguage, mentions	Detected language, @mentions in the text
awemeId, videoWebUrl, commentUrl	Video ID, video link, link that opens the video at this comment
input, scrapedAt	The input URL, read time

Sample row:

{
  "cid": "7106606555949433646",
  "text": "15?",
  "diggCount": 156,
  "replyCommentTotal": 4,
  "createTimeISO": "2022-06-07T21:03:06.000Z",
  "uniqueId": "0_.preppy._.vids._0",
  "isReply": false,
  "repliesToId": null,
  "isPinned": false,
  "authorLiked": false,
  "awemeId": "7106594312292453675",
  "videoWebUrl": "https://www.tiktok.com/@tiktok/video/7106594312292453675"
}

Input options

• postURLs: video URLs, vm.tiktok.com short links or bare video IDs.
• commentsPerPost (default 100): top-level comments per video. Replies come on top.
• maxRepliesPerComment (default 0): replies to read under each comment.
• excludePinned, oldestCommentDate (date or "7 days"), minCommentLikes: filters; skipped comments are not charged.
• maxItems: hard cap on rows across all videos. A run also stops at your Max cost per run setting.
• monitorMode + monitorStoreName: only comments not delivered before.

Pricing

Pay per event. No subscription, no rental fee.

Event	Free	Bronze	Silver	Gold	Platinum	Diamond
Comment (one comment or reply delivered)	$0.00087	$0.0007	$0.00052	$0.00015	$0.00015	$0.00015

Comments	Free	Bronze	Silver	Gold+
1,000	$0.87	$0.70	$0.52	$0.15
10,000	$8.70	$7	$5.20	$1.50
100,000	$87	$70	$52	$15

Not charged: videos that cannot be read after retries, comments removed by a filter, and comments already delivered in monitor mode.

How to run it

• Console: fill the form, press Start.
• API: POST https://api.apify.com/v2/acts/transparent_meteorite~tiktok-comments-scraper/run-sync-get-dataset-items?token=<TOKEN> with the input JSON as the body.
• Python: ApifyClient(token).actor("transparent_meteorite/tiktok-comments-scraper").call(run_input={"postURLs": ["https://www.tiktok.com/@tiktok/video/7106594312292453675"]})
• Schedules and integrations: run hourly with monitor mode on and send new comments to Google Sheets, Slack, Make, Zapier or n8n.

Limits and honest caveats

• Comments come in TikTok's own order (relevance first), not strictly by date. Use oldestCommentDate to cut old ones.
• TikTok hides some comments from logged-out visitors (filtered or restricted). Those are not returned, so the count can be a little below the number shown in the app.
• Comments on private, deleted or age-restricted videos cannot be read; those videos are reported in the log and not charged.

Common workflows and use cases

• Community management and moderation. Pull every comment under a campaign video, sort by diggCount, and answer the most-liked questions first. With excludePinned on you skip the creator's own pinned message, and with authorLiked you can see which comments the creator already acknowledged. Schedule the actor hourly with monitor mode so your team only sees comments that arrived since the last run.

• Brand and sentiment research. Collect a few thousand comments under videos that mention your product or a competitor's, export to CSV and feed text into any sentiment or topic model. The commentLanguage field lets you split results by language without running a separate detector, and mentions shows which accounts people tag in the thread.

• Creator vetting. Before paying an influencer, read the comments under their last ten sponsored posts. A healthy audience asks real questions, replies to each other and uses varied wording. A bought audience posts the same emoji or a single generic phrase. replyCommentTotal and the threading fields (isReply, repliesToId) make this easy to measure in a spreadsheet.

• Giveaway and contest picking. Set commentsPerPost high, export unique uniqueId values and draw winners. Because every row carries commentUrl, you can prove each entry by opening the video at that exact comment.

• Lead and question mining. Search your niche for videos, run this actor on the top ones and filter text for words such as "price", "where", "how much" or "link". Those commenters are people raising their hand, and profileUrl takes you to their account.

• Research and archiving. createTimeISO and scrapedAt give you a clean timeline for academic or journalistic work. Use oldestCommentDate with a value like "30 days" to keep each export focused on a single window.

Tips for cleaner, cheaper runs

• Start small. A first run with commentsPerPost at 20 shows you the exact row shape and costs a fraction of a cent, so you can wire up your pipeline before scaling to thousands.
• Use minCommentLikes to keep only comments with real engagement. Filtered comments are skipped before delivery and are never charged, so a filter lowers cost as well as noise.
• Replies cost the same as comments. If you only need top-level opinions, leave maxRepliesPerComment at 0. If you want whole conversations, set it to 5 or 10 and rebuild the thread with repliesToId.
• Set Max cost per run in the run options. The actor stops cleanly when the cap is reached and keeps everything already delivered.
• Mix video URL formats freely. Full links, vm.tiktok.com short links and bare numeric video IDs can sit in the same postURLs list.
• For recurring jobs, give each task its own monitorStoreName. Two schedules that share a store would hide each other's new comments.

Output and integrations

Results land in the default dataset and can be downloaded as JSON, CSV, Excel, XML or HTML, or streamed through the API while the run is still going. The overview table in the Output tab shows the commenter, the text, likes, reply count and the link to the comment. From there you can connect the dataset to Google Sheets, Airtable, Slack, Make, Zapier or n8n with a webhook on run success, or call the run-sync endpoint from your own code and receive rows in the HTTP response.

A typical automation: a schedule runs every hour on your ten latest videos with monitor mode on, a webhook fires when the run succeeds, and a workflow posts new comments with more than five likes into a team Slack channel along with the commentUrl. Another popular setup stores every run in a warehouse table keyed on cid, so you can track how a comment's like count grows over time.

Data quality notes

Every row is a real, publicly visible comment. Text is delivered as written, including emoji, line breaks and non-Latin scripts, encoded as UTF-8. Duplicate comments are not returned within one run. Like counts and reply counts are a snapshot at scrapedAt, so compare two runs if you need growth. Commenter display names can change at any time; uid and secUid are the stable identifiers to join on.

Who uses this actor

Social media managers use it to keep a weekly record of audience questions across every video they publish. Agencies run it for client reports, pairing comment counts and top comments with view data from a profile scraper. Researchers use it to study how conversations form and spread under viral videos, and e-commerce teams use it to find purchase-intent comments and recurring product complaints that never reach support. Because the input accepts any number of video links, one scheduled task can cover a whole content calendar, and because filtered or unreadable comments are not charged, you only ever pay for rows you actually receive.

FAQ

Do I need a TikTok account or cookies? No.

Does it get replies? Yes: set maxRepliesPerComment. Each reply is its own row with isReply: true and repliesToId pointing at the parent comment.

How fast is it? About 50 comments per request; 100 comments usually take a few seconds, and three videos are read in parallel.

Can I watch a video for new comments? Yes: turn on monitor mode and schedule the actor. Each run returns only comments it has not delivered before.

What does a failed video cost? Nothing. Only delivered comments are charged.

Use it from AI assistants (MCP)

TikTok Comments Scraper works as a tool for AI assistants through the official Apify MCP server. Add this URL to any MCP client (Claude Desktop, Claude Code, Cursor, ChatGPT connectors, VS Code):

https://mcp.apify.com/?actors=transparent_meteorite/tiktok-comments-scraper

{
  "mcpServers": {
    "apify": {
      "url": "https://mcp.apify.com/?actors=transparent_meteorite/tiktok-comments-scraper",
      "headers": { "Authorization": "Bearer <YOUR_APIFY_TOKEN>" }
    }
  }
}

Then ask, for example: "Get the top 200 comments of this TikTok video and summarise what people like and dislike."

Terms and responsible use

The actor reads only comments TikTok shows publicly to logged-out visitors. You are responsible for using the data lawfully (GDPR, CCPA, TikTok's terms). Do not use it for spam or harassment.

Changelog

• 0.2 (2026-10-28): listing refresh (title, description, README sections, console prefills). No change to output fields.

• 0.1 (2026-10-13): first release. Comments, replies, pinned and creator-liked flags, filters, monitor mode.

Support

Found a bug or need a field? Open an issue on the Issues tab with the run URL. We reply there.

Related actors

Other public-data actors from the same publisher that pair well with this one:

• TikTok Scraper: Hashtags, Search, Sounds, Places, Profiles
• TikTok Profile Scraper: Followers, Videos, Bio Email
• YouTube Comments Scraper: Comments & Replies Export

Example input

A small run that finishes quickly (this is also the prefilled input in the Apify Console). Values come from the input schema:

{
  "postURLs": [
    "https://www.tiktok.com/@tiktok/video/7106594312292453675"
  ],
  "commentsPerPost": 50,
  "maxRepliesPerComment": 0,
  "excludePinned": false,
  "maxItems": 20,
  "monitorMode": false,
  "monitorStoreName": "tiktok-comments-monitor",
  "proxyConfiguration": {
    "useApifyProxy": true
  }
}

Output sample

Each result is one dataset item (JSON, CSV, Excel or via API). Shape taken from the actor's dataset schema; field names and meanings below, not live data:

Field	Meaning
avatarThumbnail	Commenter avatar (100x100)
uniqueId	Commenter username
text	Comment text
diggCount	Likes on the comment
replyCommentTotal	Number of replies (top-level comments)
createTimeISO	Posted at, ISO 8601
isReply	True for replies
repliesToId	For replies: ID of the parent comment
isPinned	Pinned by the video creator
authorLiked	Liked by the video creator
videoWebUrl	Video URL

Alternatives to TikTok Comments Scraper

Other ways to get similar data, so you can choose what fits:

• TikTok Research API: official, restricted to approved researchers
• Apify's TikTok Scraper (clockworks/tiktok-scraper): another Apify Store actor

This actor's edge is pay-per-result pricing, no login and ready-to-use structured output. Compare on a small test run before you commit.