TikTok Scraper - Videos, Hashtags, Profiles & Transcripts

What does TikTok Scraper do?

TikTok Scraper extracts TikTok videos from tiktok.com and turns a profile, a hashtag or a video URL into structured rows: one post per row with its caption, plays, likes, comments, shares, saves, publish date, author stats, sound, video files and, as an add-on, the transcript of what is said in the video. Built for social-media analysts, influencer-marketing teams, brand monitors and researchers who need every TikTok post of an account or a topic, or only the new ones since yesterday, without scrolling the app.

TikTok Scraper - TikTok Video Scraper - TikTok Transcript Scraper - TikTok Hashtag Scraper - TikTok Profile Scraper

• Transcripts next to the stats: the spoken text of the video from TikTok's own caption files (creator captions or TikTok's speech recognition) in a transcript field, no AI cost and no separate tool - 68% of posts on a 25-post platform run had one.
• Profiles, hashtags and video URLs, one row shape: profiles (videos or reposts), hashtags and direct video or photo-post URLs, all in one run, all delivering the same fields. Words in Search queries are read as the public hashtag feed for the same words, because TikTok shows keyword search only after login; the run says so (see Limits).
• Every number the post shows: plays, likes, comments, shares, saves and reposts, plus a computed engagementRate, exact createTimeISO, duration, resolution and the cover image.
• The author on every row: username, nickname, verified badge, bio, followers, total likes and video count; on profile runs also the bio link and seller flag.
• Filters: date window, minimum likes, exclude ads and paid partnerships, exclude pinned posts, videos-only or photo-posts-only, keyword filter.
• No login, cookies or TikTok API key: paste names, tags, keywords or URLs and nothing else.
• Monitor mode for schedules: each run delivers only posts it has not delivered before; a profile checked daily costs one page.
• Export anywhere: JSON, CSV, Excel, XML, RSS; schedules, webhooks, Make, Zapier, n8n, Google Sheets, Slack; Apify API and MCP server, with a compact mode for AI agents.

What data can you get from TikTok?

🆔 Post id and URL (`id`, `webVideoUrl`)	📝 Caption and its language (`text`, `textLanguage`)	📅 Publish time (`createTime`, `createTimeISO`)
▶️ Plays (`playCount`)	❤️ Likes (`diggCount`)	💬 Comments (`commentCount`)
🔁 Shares and reposts (`shareCount`, `repostCount`)	🔖 Saves (`collectCount`)	📈 Engagement rate (`engagementRate`, percent of plays)
👤 Author: username, nickname, verified, bio, bio link, avatar (`authorMeta`)	👥 Author stats: followers, following, total likes, videos (`authorMeta.fans`, `.heart`, `.video`)	🎵 Sound: name, artist, original flag, sound id, play URL (`musicMeta`)
🎬 Video: duration, width, height, definition, format, cover, play and download URLs (`videoMeta`)	🔤 Subtitle files per language (`videoMeta.subtitleLinks`) and transcript text (`transcript`, add-on)	#️⃣ Hashtags with ids (`hashtags`) and mentions (`mentions`, `detailedMentions`)
🖼️ Photo posts: `isSlideshow`, image links (`slideshowImageLinks`)	📢 Ad and paid-partnership flags (`isAd`, `isSponsored`), pinned flag (`isPinned`)	📍 Tagged place (`poi`), attached links (`anchors`), TikTok Shop flag (`hasTikTokShopProduct`)
🏷️ Content category labels (`diversificationLabels`, `categoryType`)	🤖 AI-generated label (`isAIGC`, `aigcLabel`)	⚙️ Duet, stitch, share and comment settings

On a 25-post run on the Apify platform (profile @tiktok, #cooking, the query "cooking recipes" and one video URL) every row had the id, caption, publish time, all six engagement counts, the author's name and stats, the sound, the cover image and the video file links; 72% carried subtitle files, 68% hashtags, 44% mentions and 24% on-screen text stickers. Photo slideshows (isSlideshow, no video file) were about 1 in 100 posts across 208 captured hashtag and profile posts.

How to scrape TikTok with TikTok Scraper

1. Create a free Apify account - $5 of monthly credit, no card needed.
2. Open TikTok Scraper and click Try for free.
3. Paste usernames into Profiles, tags into Hashtags or video links into Video URLs (words in Search queries are read as the matching hashtag feed), switch on Include transcript for the spoken text, and set Max posts per profile, hashtag or search.
4. Click Start. Posts stream into the dataset as TikTok serves them, about 30 per feed page.
5. Open the Storage tab and Export as JSON, CSV or Excel - or let a schedule run it for you with Only new posts since the last run switched on.

⬇️ Input

Parameter	Type	Description
🔍 What to scrape (fill any of the four; URLs are routed by what they are)
profiles	List	Usernames or profile URLs: tiktok, @zachking, https://www.tiktok.com/@khaby.lame. Newest videos first, about 35 per feed page
hashtags	List	Hashtags or tag URLs: cooking, #fyp, https://www.tiktok.com/tag/dance
searchQueries	List	Keywords or search URLs: cooking recipes, ai tools 2026. TikTok login-walls keyword search, so a query is read as the hashtag feed for the same words (#cookingrecipes); rows carry searchQuery and searchHashtag
postURLs	List	Video or photo-post URLs, short links or bare ids: https://www.tiktok.com/@zachking/video/7095025543627705643, https://vm.tiktok.com/ZMabc123/, 7095025543627705643
🎯 Filters (applied by TikTok)
profileSection	Select	videos (the creator's own posts) or reposts (videos the creator reposted; authorMeta is the original creator). Default: videos
profileSorting	Select	latest (complete, newest-first), popular (TikTok's Popular order), oldest (oldest video first). Default: latest
✂️ Refine results
oldestPostDate	String	Only posts published after: 2026-09-01, 7 days, 3 weeks, or a day count (1 = today). On a latest-first profile the run stops at the first page older than this
newestPostDate	String	Only posts published before: 2026-06-30, 30 days
leastDiggs	Integer	Minimum likes, e.g. 10000. Default 0
excludePinnedPosts	Boolean	Skip posts pinned to the top of a profile. Default false
excludeAds	Boolean	Skip ads and paid partnerships (isAd, isSponsored). Default false
postType	Select	all, video (skip photo slideshows) or photo (slideshows only). Default all
keywordFilter	String	Keep posts whose caption, hashtags, on-screen text or author name contain the word(s), comma = any: giveaway, recipe, meal prep
➕ Extra data
includeTranscript	Boolean	Adds transcript, transcriptLanguage, transcriptSource from TikTok's caption file - charged as the transcript event only for posts that receive one. Default false
🔁 Monitoring
monitorMode	Boolean	Deliver only posts not delivered by previous runs with the same monitorStoreName. Pair with a daily schedule. Default false
monitorStoreName	String	Key-value store that keeps the per-target memory. Default tiktok-posts-monitor
📦 Output
resultsPerPage	Integer	Max posts per profile, hashtag or search after filters. 0 = everything TikTok serves. Default 30 (prefill 10)
maxItems	Integer	Hard cap for the whole run. 0 = no total cap. Default 0
compact	Boolean	Core fields only (id, text, date, URL, author, counts, engagement rate, duration, hashtags, transcript), sized for AI agents and MCP. Default false
⚙️ Advanced
proxyConfiguration	Proxy	Default: Apify Proxy, RESIDENTIAL group. With this default, feeds are read through Apify datacenter proxies first and through the residential group only when TikTok refuses those; video pages and subtitles go through datacenter proxies automatically. Custom proxies are used as given

Input JSON

{
  "profiles": ["tiktok", "https://www.tiktok.com/@zachking"],
  "hashtags": ["cooking"],
  "searchQueries": ["cooking recipes"],
  "postURLs": ["https://www.tiktok.com/@zachking/video/7095025543627705643"],
  "resultsPerPage": 30,
  "oldestPostDate": "30 days",
  "excludeAds": true,
  "includeTranscript": true
}

Give it the name, not the search for the name

A profile is zachking or @zachking or the profile URL; a hashtag is cooking or #cooking; only free text goes into Search queries, where it is read as the hashtag feed for the same words. Do: profiles: ["zachking"] · Don't: searchQueries: ["zachking"] (that reads the #zachking hashtag feed: other people's videos tagged with the name, not his own posts).

Use latest-first for anything complete, scheduled or date-bounded

TikTok's profile feed is newest-first, so oldestPostDate and monitorMode can stop reading as soon as a page is older than the floor or the last run's newest post. Hashtag feeds (and search queries, which are served as hashtag feeds) are TikTok's ranking: the date window still applies, but the whole page is read first, so set resultsPerPage to what you actually need.

A video URL is the cheapest way to refresh known posts

Each URL in postURLs is one HTML page fetched through a datacenter proxy, no browser session, and returns the same row shape, plus locationCreated (the country the post was made in), which TikTok only publishes on the video page. Short links (vm.tiktok.com/..., tiktok.com/t/...) are resolved first.

⬆️ Output

Every TikTok post becomes one dataset item. Here is the table view of a platform run (profile @tiktok, #cooking and one video URL, transcript add-on on):

And one row as JSON (long URL fields shortened):

{
  "id": "7095025543627705643",
  "text": "#answer to @spidermark790",
  "textLanguage": "en",
  "createTimeISO": "2022-05-07T15:22:45.000Z",
  "locationCreated": "US",
  "authorMeta": { "name": "zachking", "nickName": "Zach King", "verified": true, "fans": 86900000, "heart": 1300000000, "video": 618, "profileUrl": "https://www.tiktok.com/@zachking" },
  "musicMeta": { "musicName": "original sound", "musicAuthor": "Zach King", "musicOriginal": true, "musicId": "7095025525487962926", "duration": 18 },
  "webVideoUrl": "https://www.tiktok.com/@zachking/video/7095025543627705643",
  "videoMeta": { "height": 1024, "width": 576, "duration": 18, "definition": "540p", "format": "mp4", "coverUrl": "https://p16-sign.tiktokcdn-us.com/...", "downloadAddr": "https://v16-webapp-prime.tiktok.com/...", "subtitleLinks": [{ "language": "eng-US", "source": "ASR", "downloadLink": "https://v16-webapp.tiktok.com/..." }] },
  "playCount": 27100000, "diggCount": 1400000, "commentCount": 13200, "shareCount": 16600, "collectCount": 47672, "repostCount": 0,
  "engagementRate": 5.45,
  "hashtags": [{ "id": "92758", "name": "answer", "title": "", "cover": "" }],
  "mentions": ["@spidermark790"],
  "isAd": false, "isSponsored": false, "isSlideshow": false, "isPinned": false,
  "diversificationLabels": ["Professional Special Effects"],
  "transcript": "you like my costume? so cool how are you doing that...",
  "transcriptLanguage": "eng-US",
  "submittedVideoUrl": "https://www.tiktok.com/@zachking/video/7095025543627705643",
  "scrapedAt": "2026-10-03T07:25:11.450Z"
}

Fields

Field	Type	What it is	Fill on a 25-post platform run
id, webVideoUrl	string	Post id and canonical URL (/photo/ for slideshows)	100%
text, textLanguage	string	Caption and TikTok's language code	100%
createTime, createTimeISO	integer, string	Publish time, Unix seconds and ISO 8601	100%
playCount, diggCount, commentCount, shareCount, collectCount, repostCount	integer	Plays, likes, comments, shares, saves, reposts	100%
engagementRate	number	(likes + comments + shares + saves) / plays, in percent	100%
authorMeta	object	id, name, nickName, profileUrl, verified, signature, avatar, secUid, privateAccount, following, friends, fans, heart, video, digg; on profile runs also bioLink, commerceUserInfo, ttSeller, region, roomId	100% (bioLink 32%: the profile rows)
musicMeta	object	musicName, musicAuthor, musicOriginal, musicAlbum, musicId, duration, playUrl, coverMediumUrl, isCopyrighted, isCommerceMusic	100%
videoMeta	object	height, width, duration, definition, format, bitrate, codecType, size, coverUrl, originalCoverUrl, dynamicCoverUrl, playAddr, downloadAddr, subtitleLinks[]	100% (file links on every video; photo posts have none)
mediaUrls	array	Direct media URLs (expire after a few hours)	100%
hashtags, mentions, detailedMentions	array	Hashtags with ids; mentioned accounts with id, secUid and profile URL	68% / 44%
isSlideshow, slideshowImageLinks	boolean, array	Photo posts and their image links	0% in this run; about 1% of 208 captured posts
isAd, isSponsored, isPinned	boolean	TikTok's ad flag, paid-partnership flag, pinned-to-profile flag	flags; 1 of 25 here, 15% of brand-account posts on a 210-post sample
subtitleLanguages, transcript, transcriptLanguage, transcriptSource	array, string	Caption languages; the transcript text (add-on)	subtitle files 72%, transcript 68%
poi, anchors, hasTikTokShopProduct	object, array, boolean	Tagged place; attached links (CapCut templates, shop products); shop flag	0% / 24% / 0% here (poi 3% of 208 captured posts)
diversificationLabels, categoryType, isAIGC, aigcLabel	array, integer, boolean, string	TikTok's content category labels and AI-generated label	labels 4% (video URL rows only), category 100%
duetEnabled, stitchEnabled, shareEnabled, commentsDisabled	boolean	Interaction settings	100%
input, fromProfileSection, searchQuery, searchHashtag, submittedVideoUrl	string	Which input produced the row	100%

How can I use TikTok data?

• 🎯 Influencer-marketing teams - pull the last 30 videos of 50 creators, rank by engagementRate and fans, and shortlist who actually moves an audience.
• 📊 Social-media managers and agencies - track a brand account and its competitors daily with monitorMode; every new video lands in a Google Sheet with its first-day plays.
• 🔎 Trend and hashtag research - read #cooking or #fyp each morning, keep posts with leastDiggs: 10000, and see which sounds (musicMeta.musicName) and formats are rising.
• 🛍️ Brands and e-commerce - read the hashtag of your product or brand, filter paid partnerships out with excludeAds, and read what creators say about it in transcript.
• 🤖 AI and research teams - feed text, transcript and hashtags to an LLM for sentiment or topic analysis; compact keeps a thousand posts inside a context window.
• 📈 Content strategists - compare postType video vs photo, duration vs engagementRate, posting cadence from createTimeISO, week over week on a schedule.

How much does it cost to scrape TikTok?

TikTok Scraper is pay-per-event (the Pricing tab always shows the live rates):

Event	Price	When it is charged
feed-scraped	$0.005 per profile, hashtag or search query	Once per target per run, when TikTok serves its feed (opening it in a browser session). Never for video URLs
post-scraped	$0.00028 per post ($0.28 per 1,000)	Each post written to the dataset after your filters
transcript	$0.001 per transcript ($1.00 per 1,000)	Optional add-on; only posts that receive a transcript

There is no start fee. Once TikTok serves a feed, its fee applies even if your filters remove every post or, in monitor mode, nothing is new: opening and checking the feed is the service.

What one profile or hashtag costs:

Posts	Feed fee	Posts	Total	With a transcript on every post
10	$0.005	$0.0028	$0.0078	$0.0178
100	$0.005	$0.028	$0.033	$0.133
1,000	$0.005	$0.28	$0.285	$1.285

Worked examples:

• 1 profile × 30 posts = $0.005 + 30 × $0.00028 = $0.005 + $0.0084 = $0.0134
• 1 profile × 30 posts with transcripts (about 7 in 10 videos have a caption file, so about 21 transcripts) = $0.0134 + 21 × $0.001 = $0.0344
• 20 video URLs = 20 × $0.00028 = $0.0056 (no feed fee)
• Daily monitor of 10 creators with 2 new posts each = 10 × $0.005 + 20 × $0.00028 = $0.05 + $0.0056 = $0.0556; a day with nothing new costs $0.05

There is nothing else to pay: Apify platform usage (compute, residential proxy) is included in these prices. Measured on the platform on 2026-10-07 (default 2 GB of memory): a 25-post profile run took 23 seconds end to end, and 125 posts across two profiles, two hashtags and a search query took 35 seconds; a profile page delivers about 35 posts and a hashtag page about 30.

Integrations, API, scheduling & MCP

• Schedules: run TikTok Scraper hourly, daily or weekly from the Schedules tab; combine with monitorMode to get only new posts, and with oldestPostDate: "2" to re-read yesterday's and today's posts with fresh counts.
• Webhooks: fire a URL when a run finishes - push rows to your CRM, Slack or a sheet.
• No-code: Make, Zapier, n8n, Google Sheets, Airtable, Slack via Apify integrations.
• API: call it from Node.js or Python with apify-client - code on the API tab and the Apify API reference.
• MCP / AI agents: connect Claude, Cursor or any MCP client through the MCP tab and ask for TikTok videos in plain language; switch on compact to keep the answer small.

More TopApi scrapers

⭐ Google Maps Reviews Scraper - every review of a place, newest first, with owner replies and exact dates

🎵 TikTok Scraper - this actor: posts by profile, hashtag, search or URL

🧭 All TopApi social-media and lead-gen actors - Instagram, YouTube Shorts, Facebook and X scrapers are next on this account

Limits & empty runs

• Keyword search is a login feature on TikTok's web app. To the signed-out sessions this actor uses, the search page shows "Log in to search for popular content" and never loads results (measured on every platform session on 2026-10-03). The actor never logs in, so a search query is served as the public hashtag feed for the same words (cooking recipes -> #cookingrecipes), every row carries searchQuery and searchHashtag, and the status message says how many queries were served that way. A query with no hashtag equivalent delivers nothing.
• Feeds are read in a real browser session. The run opens TikTok's web app once, then requests each feed page from TikTok the way the app does, through Apify datacenter proxies (residential when TikTok refuses those). If TikTok still serves no feed, that target is read from TikTok's own page in the browser, through fresh residential sessions (up to three per target), and otherwise reported as "got no data from TikTok".
• Posts per feed page: TikTok serves about 35 posts per profile page and about 30 per hashtag page; a profile with 1,500 videos is about 45 pages. resultsPerPage and oldestPostDate are the cost controls.
• Cut-down hashtag feeds: TikTok serves some visitors a short hashtag feed with no further pages. Measured on 2026-10-07: #fitness answered 10-11 posts to 6 of 9 fresh sessions (and 30 per page, 100+ in all, to the others); #fyp and #foryou answered 1-5 posts every time. A hashtag that comes back short is read once more on a fresh session; what TikTok serves then is what the run delivers.
• Speed: the first feed page arrives about 5-20 seconds after the run starts (browser start and TikTok's web app booting, once per run); each further page takes 1-3 seconds. A target that needs the fallback takes from 30 seconds to two minutes more.
• Sorted profile feeds: popular follows TikTok's Popular order and oldest starts at the account's first video. If a profile has to be read from TikTok's own page and that page serves no sorted feed, the run reads it latest-first and logs it.
• Reposts: only profiles that show a Reposts tab; a profile without one returns no rows for profileSection: "reposts".
• Empty dataset: means no matches after your filters, a private, removed or non-existent target, or TikTok served no data to the session (the status message names which). A feed TikTok did serve still pays its feed fee.
• Not available: comment text and follower lists (separate products), email or phone of the author (TikTok publishes none), view history over time (each run is a snapshot). Media URLs expire after a few hours; re-run the video URL to refresh them.

FAQ

Can TikTok Scraper give me the transcript of a TikTok video?

Yes, with includeTranscript: true: it fetches TikTok's own caption file (the creator's captions or TikTok's automatic speech recognition) and delivers the plain text in transcript, with the language and source, on the same row as the post's stats. Videos without a caption file return null.

How many TikTok posts can TikTok Scraper collect in one run?

Everything TikTok serves: resultsPerPage: 0 walks a profile to its oldest video (a 1,500-video account is about 45 feed pages, roughly a minute). Hashtag feeds (and search queries, served as hashtag feeds) are TikTok's ranking and go as deep as TikTok pages them; on tests #cooking kept serving 28-30 new posts per page, while some hashtags are cut short for some visitors (see Limits).

Can I run TikTok Scraper on a schedule to get new TikTok videos automatically?

Yes. Save your input as a task, open Schedules, pick the cadence, and switch on monitorMode. Each run delivers only posts not delivered before and stops at the last run's newest post on profile feeds, so a daily check of a big account costs one page. A run that finds nothing new delivers nothing and bills only the $0.005 feed fee per profile, hashtag or query checked.

Can I integrate TikTok Scraper with other apps?

Yes - Make, Zapier, n8n, Google Sheets, Slack and more via Apify integrations, or use webhooks to trigger an action when a run finishes.

Can I use TikTok Scraper with the Apify API?

Yes. The API lets you start runs, fetch datasets and manage schedules. Use apify-client for Node.js or Python; the API tab has ready-to-paste code.

Can I use TikTok Scraper through an MCP server?

Yes. Open the MCP tab, connect your MCP client (Claude Desktop, Cursor, LibreChat...), and ask for TikTok videos of a creator or hashtag in plain language. Set compact: true so a hundred posts fit the model's context.

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

No. TikTok Scraper reads only what TikTok shows a signed-out visitor; you pass names, hashtags, keywords or URLs and nothing else. That is also why keyword search is served as a hashtag feed: TikTok's search results page asks visitors to log in, and this actor never does.

Why did my run return no TikTok videos?

In order of frequency on test runs: (1) the target is private, removed or misspelled - the status message says "not found" or "private"; (2) your filters removed everything - the run summary lists how many posts were read and why they were dropped; (3) monitorMode found nothing new since the last run; (4) a search query had no hashtag equivalent (TikTok login-walls keyword search, so queries are served as hashtag feeds); (5) TikTok served no feed data to the run's sessions, the browser fallback included - retry in a few minutes or use postURLs for posts you already know. In (2) and (3) TikTok did serve the feed, so its feed fee applies.

Does TikTok Scraper download the video files?

It delivers the direct media URLs (videoMeta.downloadAddr, mediaUrls) and the cover images; the files themselves are not stored. The URLs expire after a few hours, so download them right after the run or re-run the video URL.

How is TikTok Scraper different from other TikTok scrapers?

It puts the transcript text on the same row as the post, from TikTok's own caption files, and delivers the same post fields as the lane's leading TikTok Scraper plus engagementRate, content labels, AI-generated flags, tagged places and attached links; it filters by date, likes, ads, pinned posts, type and keyword, and ships monitor mode for scheduled runs. On the 32-post test run every row carried all six engagement counts and the author's stats.

Is it legal to scrape TikTok data?

TikTok Scraper extracts only publicly available TikTok posts; it does not log in or collect private data. Results may still contain personal data (creator names, captions), which is regulated by laws such as GDPR. Make sure you have a legitimate purpose before you use it.

Your feedback

We keep improving TikTok Scraper. Found a bug or missing a field? Open an issue on the Issues tab - we answer within two working days.