TikTok Video Scraper With Engagement Data

Scrape TikTok videos by direct URL and get, in addition to the full video/author/music metadata, a computed engagement-analytics field group on every row -- engagement rate, like/view and comment/view ratios, a virality score, a duration bucket, and a batch-relative engagement percentile rank -- plus optional filters to drop low-performing videos from the output.

What it does

1. Fetches each submitted TikTok video URL and extracts the video's full public metadata (author, music, stats, hashtags, mentions, media URLs, moderation/content flags).
2. Optionally follows each video to its related-videos feed (scrapeRelatedVideos).
3. Computes a set of derived analytics fields for every successfully scraped video.
4. Optionally drops videos below a minimum engagement rate and/or virality score.
5. Every row -- including ones that failed to scrape -- is pushed with the exact same set of columns, so a failure shows up as a clearly labeled error row instead of silently disappearing or rendering as a blank/short row.

Input

Field	Type	Default	Notes
postURLs (required)	array of strings	1 sample URL	Direct TikTok video URLs to scrape.
scrapeRelatedVideos	boolean	false	When enabled, also scrapes each video's related-videos feed.
resultsPerPage	integer (1-20)	10	Related videos to fetch per post, when scrapeRelatedVideos is enabled. TikTok's own related-video surface tops out around 12-16 videos per post, so values above that will not return more rows.
minEngagementRate	number	0 (disabled)	Drops videos whose engagementRate is below this value. 0 disables the filter.
minViralityScore	number	0 (disabled)	Drops videos whose viralityScore is below this value. 0 disables the filter.

Rows that failed to scrape are never removed by either filter -- filters only apply to rows with real engagement data.

Example input:

{
  "postURLs": ["https://www.tiktok.com/@apifyoffice/video/7200360993149553925"],
  "scrapeRelatedVideos": false,
  "resultsPerPage": 10,
  "minEngagementRate": 0,
  "minViralityScore": 0
}

Analytics fields (computed, not from TikTok)

• engagementRate = (likes + shares + comments + saves) / views
• likeToViewRatio = likes / views
• commentToViewRatio = comments / views
• viralityScore = engagementRate / age of the video in days (from createTimeISO to the time the run executed) -- rewards videos that gathered engagement quickly.
• durationBucket = short (<15s) / medium (15-60s) / long (>60s), from videoMeta.duration.
• engagementPercentileRank = this video's engagementRate percentile rank (0-100) relative to every other successfully scraped video in the same run.

All analytics fields are null on a row that failed to scrape, or when playCount is 0 (a rate can't be computed against zero views).

Output

Every dataset row shares the same ~70-field shape, whether the video scraped successfully or not. On failure, every metadata field is null/empty and error explains what happened; on success, error is null. Example of a successful row (values illustrative):

{
  "id": "7200360993149553925",
  "text": "Example caption #example",
  "createTimeISO": "2023-05-15T12:00:00.000Z",
  "authorMeta": {
    "name": "apifyoffice",
    "verified": false,
    "fans": 1234
  },
  "diggCount": 5400,
  "shareCount": 210,
  "playCount": 98000,
  "commentCount": 340,
  "collectCount": 180,
  "mediaUrls": ["https://v16-webapp.tiktok.com/....mp4"],
  "mentions": ["@someuser"],
  "hashtags": [{ "name": "example" }],
  "isAigc": false,
  "diversificationLabels": ["Comedy"],
  "duetEnabled": true,
  "stitchEnabled": true,
  "shareEnabled": true,
  "engagementRate": 0.0625,
  "likeToViewRatio": 0.0551,
  "commentToViewRatio": 0.0035,
  "viralityScore": 0.00021,
  "durationBucket": "medium (15-60s)",
  "engagementPercentileRank": 87.5,
  "webVideoUrl": "https://www.tiktok.com/@apifyoffice/video/7200360993149553925",
  "error": null
}

mediaUrls contains direct CDN video URLs already returned by TikTok's own page data -- this actor never downloads or stores video files.

Notes

• Uses TikTok's public video-page metadata; no login, cookies, or private-account access.
• No proxy configuration is included in this actor.
• commentsDatasetUrl and authorMeta.followDatasetUrl are always null -- this actor does not fetch comments or follower lists.