TikTok Video Scraper: Post Stats, Transcripts

TikTok Video Scraper: Post Stats, Transcripts

TikTok Video Scraper: paste TikTok video URLs, photo slideshow URLs, short links or post ids and get one row per link, in the order you gave them, with no login, no cookies and no API key. Each row carries the play, like, comment, share, save and repost counts, the caption with its hashtags and mentions, the music, the author and the author's follower and like counts at the time of the run, TikTok's own content categories and related searches, the media URLs with the time they expire, and the transcript of the video from TikTok's automatic caption track, inline and free.

You pay per post delivered. A link that could not be served gets a free status row in its place that says why (not found, private, blocked, invalid, duplicate, filtered), and the run summary is free.

Quick start

1. Open the Actor and paste your links into TikTok post URLs, short links or post ids, one per line. Full www.tiktok.com/@user/video/... and /photo/... URLs, vm.tiktok.com and vt.tiktok.com short links, m.tiktok.com/v/....html links and bare post ids all work.
2. Leave Include the transcript on to get the spoken text of each video at no extra cost.
3. Set Maximum posts if you want a cap (0 means every link).
4. Run it. Download the Posts view as JSON, CSV or Excel, or read the dataset through the API. Row 0 is your first link, row 1 your second, and so on.
5. For a list you re-check on a schedule, give the run a Memory name: posts already delivered under that name are not fetched or charged again.

What you get

Every row has position (the index of your link), status and input (your link as given). A delivered post (type: "post", status: "ok") has the fields below. Fill rates are measured on the Apify platform on 174 live posts from brand and creator accounts:

Field	Filled in the 174 post sample (run 98WgGaP12B6r8TfZf)
playCount, diggCount, commentCount, shareCount, collectCount	174 of 174
authorMeta.fans, authorMeta.heart (author at post time)	174 of 174
categoryType	174 of 174
textLanguage	174 of 174
mediaExpiresAt	174 of 174
commentsEnabled	174 of 174
videoMeta.downloadAddr, vqScore, loudness, dynamicCoverUrl	158 of 158 videos
musicMeta.playUrl	164 of 174
suggestedWords (related searches)	145 of 174
diversificationLabels (content categories)	140 of 174
hashtags	111 of 174
transcript	82 of 158 videos
mentions	47 of 174
musicMeta.dspLinks	54 of 174
isAd true	27 of 174
poi (tagged place)	11 of 174

Field	What it is
id, webVideoUrl, resolvedUrl	TikTok's post id (a string), the canonical URL (/photo/ for slideshows) and, for short links, where the link led
text, textLanguage, hashtags, mentions, detailedMentions	caption, its language, hashtags with ids, and mentioned accounts with their ids
createTime, createTimeISO	when the post was published
playCount, diggCount, commentCount, shareCount, collectCount, repostCount	engagement counts from TikTok's statsV2 block (see the note below)
authorName, authorMeta	the author as TikTok serves it with the post (never taken from your URL), with follower, following, like and video counts, verified, bio, bio link and account settings
musicMeta	song id, title, artist, original or not, copyright and commerce flags, the audio URL and streaming links when TikTok has them
videoMeta	duration, size, cover URLs, downloadAddr, the largest rendition (or every rendition), TikTok's video quality score, loudness, caption track links and TikTok's caption flags
imagePost, slideshowImageLinks, isSlideshow	for photo posts: every image with its size, a second host and its expiry
diversificationLabels, categoryType, suggestedWords	TikTok's content categories and the related searches TikTok shows under the post
isAd, isAigc, commentsEnabled, duetEnabled, stitchEnabled, penaltyStatus, warnInfo	flags TikTok sets on the post
transcript, transcriptSegments, transcriptLanguage, transcriptSource, transcriptReason	the caption track as plain text and as timed segments in seconds, or the reason there is none
mediaUrls, mediaExpiresAt	the signed media URLs and the earliest moment one of them stops working

About the counts. On its web page TikTok prints play, like, comment and share counts to 4 significant digits once they reach 10,000 (212,000, not 212,113); below 10,000 they are exact. Saves (collectCount) are exact at every size. The Actor returns TikTok's figures as numbers and null, never 0, when TikTok does not give one.

About the transcript. It is TikTok's own automatic speech to text track (transcriptSource: "ASR"), downloaded and returned inline. TikTok had a track for 82 of 158 videos (52 percent) in the sample above and for 13 of 36 videos (36 percent) in a sample of hashtag feed posts. Every other post says why in transcriptReason: no_captions (TikTok made no track, 71 videos in the sample, mostly music without speech), captions_disabled (captions switched off, 5 videos, with TikTok's own code in transcriptReasonDetail), slideshow (photo posts have no speech track), download_failed or disabled_by_input. Set Transcript language to prefer a language when TikTok offers more than one track.

Status rows (type: "status", free) sit in the same place as the link they answer:

status	Meaning
not_found	TikTok says the post does not exist. TikTok answers a removed post and an id that never existed the same way, so the message says "removed or never existed". An unknown or expired short link also ends here; a short link TikTok refused on both the platform IP and the proxy is a blocked row instead.
private, removed, unavailable	TikTok answered with another code; statusCode and statusMsg echo it. A backend timeout on TikTok's side (statusCode 100004, seen once in 174 proxied loads) is fetched twice more before it becomes unavailable
blocked	TikTok did not serve the page after every retry on fresh exits
invalid	not a TikTok post link or id (a profile or hashtag URL, for example)
duplicate	the same post appeared earlier in your list; duplicateOf gives that position
already_known	delivered before under your memory name
filtered	removed by a free filter (excludeAds, slideshowsOnly)
skipped	not fetched because the run stopped first (maxPosts reached or your spending limit); stoppedBy says which
failed	an unexpected error; the message says what

The last row is the run summary (type: "summary") with counts per status, transcript outcomes, request and retry counts, the stop reason and billing counters.

How much does it cost to scrape TikTok videos?

Event	FREE and BRONZE	SILVER	GOLD	PLATINUM	DIAMOND	When it is charged
Post	$0.50 per 1,000	$0.45	$0.40	$0.40	$0.28	once per post delivered with status ok
Actor start	$0.002 per run per GB of memory	same	same	same	same	once per run; the default 512 MB run pays it once

The Store card shows "$0.50 / 1,000" on the Post row: one post costs a twentieth of a cent. A list of 1,000 links costs $0.50 plus $0.002 for the start. The transcript, the filters, the memory, status rows and the summary cost nothing. Measured on Apify, a run of 174 posts took 44 seconds, about 0.25 seconds per post with 4 parallel requests (about 230 posts per minute).

Input

Field	Type	Notes
postUrls	array	post URLs, short links or post ids; one row per value in your order
maxPosts	integer	stop after this many posts; 0 = no cap
includeTranscript	boolean	the inline transcript, free, on by default
transcriptLanguage	string	preferred caption language such as en or es
includeRenditions	boolean	every video rendition instead of the largest one
excludeAds, slideshowsOnly	boolean	free filters
memoryName, resetMemory	string, boolean	pay once per post across scheduled runs
tryDirectFirst	boolean	first request from the Apify platform IP, retries through the proxy
proxyConfiguration	object	the retry proxy; Apify datacenter proxy by default
maxConcurrency	integer	1 to 8 parallel requests, 4 by default

{
  "postUrls": [
    "https://www.tiktok.com/@nike/video/7689162344139640077",
    "https://www.tiktok.com/@nike/video/7401458428457127211"
  ],
  "maxPosts": 10,
  "includeTranscript": true
}

On the Apify platform the first request for each post went straight from the platform IP and was served on the first try for 174 of 174 in each of five runs; with every request through the datacenter proxy 508 of 522 (three runs of 174) were served on the first try and the runs took 58 to 160 seconds instead of 44. No residential proxy is used.

FAQ

Do I need a TikTok account, cookies or an API key?

No. Every post page is read the way a signed out visitor reads it, over plain HTTP: the first request goes from the Apify platform IP and any retry goes through a datacenter proxy. Nothing in the input asks for a credential.

Does it keep the order of my URLs?

Yes. The dataset carries exactly one row per URL in the order you gave it, with a position field, and a URL that could not be served gets a free status row in the same place saying why (not found, removed, private, blocked or invalid).

Does it return the transcript of a TikTok video?

When TikTok has an automatic caption track for the post, the Actor downloads it and returns the text inline plus timed segments, at no extra charge. Posts without a track carry a transcriptReason instead of an empty string.

Does it return comments?

No. Comment text is served only after a slide puzzle and a login on the web, so this Actor returns the comment count and whether comments are enabled, not the comment bodies.

Are the view and like counts exact?

Saves are exact at every size. Plays, likes, comments and shares are exact below 10,000; above that TikTok's page gives them to 4 significant digits (212,000) and the Actor returns TikTok's figure. The author's follower and like counts on a post are rounded the same way.

Can I download the video or the cover?

The row carries the signed media URLs TikTok serves (video, covers, music) and a mediaExpiresAt timestamp parsed from them. Fetch what you want to keep before that time; the Actor does not store files.

Why does this Actor run with limited permissions?

Least privilege. It reads and writes only its own run storages, and the optional cross run memory is a named key value store it creates itself.

What this Actor does NOT do

• It does not return comment text. TikTok serves comments only after a slide puzzle and a login on the web; rows carry commentCount and commentsEnabled instead.
• It does not list a profile's posts or a hashtag's posts. Give it post links; the sibling Actors below cover profiles and hashtags.
• It does not store videos or images. Media URLs are signed by TikTok and expire, about 6 hours after the run for video and music and about 47 hours for slideshow images; mediaExpiresAt gives the exact time. Download what you want to keep before then.
• It does not transcribe audio itself. The transcript is TikTok's own caption track, so a post without one has no transcript.
• It does not tell a removed post from one that never existed: TikTok answers both the same way.

API example

curl -X POST "https://api.apify.com/v2/acts/automation_craft~tiktok-video-scraper/run-sync-get-dataset-items?token=YOUR_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"postUrls":["https://www.tiktok.com/@nike/video/7689162344139640077"],"includeTranscript":true}'

Rows are about 11 KB each; the largest in the sample was 21 KB (a 9 image slideshow).

More data tools by Automation Craft

• TikTok Profile Scraper: Exact Stats, No Login
• TikTok Hashtag Scraper: Posts, Free Date Filter
• TikTok Ads Library Scraper: EU Ads, No Login
• Google Trends Scraper - Compare and Trending Now
• Meta Ad Library Scraper - All Placements, Filters
• Website Screenshot Scraper: Full Page, Bulk