TikTok Comments Scraper by Profile

Provide a TikTok username and get every comment across all their videos.

1. You input a profile (or direct video URLs)
2. All videos are discovered automatically
3. Every comment and reply is extracted — no cap

What makes this different

Most TikTok comment scrapers only accept video URLs as input, meaning you need to find each video URL yourself. This actor accepts a profile username and handles the rest — it discovers all videos from the profile and scrapes comments across all of them in one run.

Feature	This actor	Typical alternatives
Input a profile username	Yes	No (video URLs only)
Auto-discover all videos	Yes	Manual
Comment limit	None	Often capped at 5,000
Reply threads	Full depth	Varies
Image comments	Extracted	Often skipped

Input

Field	Type	Default	Description
profileUrls	string[]	[]	TikTok profile URLs or usernames (e.g. @kingconnah or https://www.tiktok.com/@kingconnah)
videoUrls	string[]	[]	Direct video URLs — combined with any videos discovered from profiles
maxCommentsPerVideo	integer	0 (all)	Maximum comments per video. 0 = no limit
includeReplies	boolean	false	Fetch reply threads for each comment

Scrape all comments from a profile

{
    "profileUrls": ["kingconnah"],
    "includeReplies": true
}

Scrape specific videos

{
    "videoUrls": [
        "https://www.tiktok.com/@username/video/7622233769231142165"
    ],
    "maxCommentsPerVideo": 1000,
    "includeReplies": true
}

Both — profile + extra videos

{
    "profileUrls": ["kingconnah"],
    "videoUrls": [
        "https://www.tiktok.com/@other/video/7622233769231142165"
    ]
}

Output

Each comment is one flat row in the dataset. Replies are separate rows linked to their parent via parent_comment_id.

Output fields

Field	Type	Description
video_url	string	The video this comment belongs to
comment_id	string	Unique comment ID
parent_comment_id	string	Parent comment ID (empty for top-level)
reply_to_reply_id	string	ID of the specific reply being responded to within a thread
depth	integer	0 = top-level, 1 = reply
comment	string	Comment text (empty for image-only comments)
comment_image_url	string	Image URL for image-only comments
likes	integer	Like count
reply_count	integer	Number of replies
created_at	string	ISO 8601 timestamp (UTC)
author_username	string	Commenter's @username
author_display_name	string	Commenter's display name
author_id	string	TikTok numeric user ID
reply_to_username	string	Username being replied to
liked_by_video_creator	boolean	Creator liked this comment
pinned	boolean	Creator pinned this comment
purchase_intent	boolean	TikTok's flag for buying intent (e.g. "Where can I buy this?")
comment_language	string	Language code ("un" = undetermined)
share_url	string	Direct share link

Example output

Top-level comment:

{
    "video_url": "https://www.tiktok.com/@txmfitt/video/7613032455011945748",
    "comment_id": "7616086972566078230",
    "parent_comment_id": "",
    "reply_to_reply_id": "",
    "depth": 0,
    "comment": "how many did you do in an hour? I once did 500",
    "comment_image_url": "",
    "likes": 62,
    "reply_count": 3,
    "created_at": "2026-03-11T19:47:22+00:00",
    "author_username": "alanas698",
    "author_display_name": "2027",
    "author_id": "6765931488842613766",
    "reply_to_username": "",
    "liked_by_video_creator": true,
    "pinned": false,
    "purchase_intent": false,
    "comment_language": "en",
    "share_url": "https://m.tiktok.com/v/7613032455011945748.html?..."
}

Reply:

{
    "video_url": "https://www.tiktok.com/@txmfitt/video/7613032455011945748",
    "comment_id": "7617031854978024193",
    "parent_comment_id": "7616086972566078230",
    "reply_to_reply_id": "",
    "depth": 1,
    "comment": "i did 1,017",
    "comment_image_url": "",
    "likes": 45,
    "reply_count": 0,
    "created_at": "2026-03-14T08:54:01+00:00",
    "author_username": "txmfitt",
    "author_display_name": "Toby",
    "author_id": "7193948041925403654",
    "reply_to_username": "",
    "liked_by_video_creator": false,
    "pinned": false,
    "purchase_intent": false,
    "comment_language": "en",
    "share_url": ""
}

Image-only comment:

{
    "video_url": "https://www.tiktok.com/@txmfitt/video/7613032455011945748",
    "comment_id": "7615909607940588310",
    "parent_comment_id": "7614996613496390408",
    "reply_to_reply_id": "",
    "depth": 1,
    "comment": "",
    "comment_image_url": "https://p16-comment-sign-no.tiktokcdn-eu.com/tos-no1a-i-jj85edgx6n-no/b5efb28ad8934663aa1831ef422142d7~tplv-jj85edgx6n-image-origin.image?...",
    "likes": 4,
    "reply_count": 0,
    "created_at": "2026-03-11T08:19:24+00:00",
    "author_username": "unicukornis",
    "author_display_name": "Unicukornis",
    "author_id": "6845319934780654598",
    "reply_to_username": "",
    "liked_by_video_creator": false,
    "pinned": false,
    "purchase_intent": false,
    "comment_language": "un",
    "share_url": "https://m.tiktok.com/v/7613032455011945748.html?..."
}

Understanding threading

For most use cases, parent_comment_id + depth is all you need:

• Top-level comments: depth == 0
• Find replies: filter by parent_comment_id
• Full threads: group by parent_comment_id, sort by created_at

The reply_to_reply_id field shows which specific reply within a thread someone responded to — useful for conversation analysis.

FAQ

How does the comment limit work with replies enabled?

When includeReplies is on, both comments and replies count toward maxCommentsPerVideo. The actor processes comments page by page — for each page, it first extracts the top-level comments, then fetches their replies before moving to the next page. This means the final output is a mix of comments and replies. If you need all top-level comments first, run once without replies, then run again with replies enabled.

What happens if a proxy gets blocked?

The actor automatically retries with a fresh proxy. You'll see a "rotating proxy" warning in the logs but scraping continues from where it left off. No data is lost.

Are there duplicate comments in the output?

No. Every comment and reply is deduplicated by ID before being added to the dataset. You only pay for unique results.