TikTok Posts Search Scraper

Apify actor that searches TikTok posts by keyword/hashtag and exports them as structured data — no login, no cookies, no session to maintain.

TikTok content is viewable logged-out; the only barrier is its anti-bot request signing. This actor drives a real Chrome (Playwright) so TikTok's own JavaScript signs its API requests, then intercepts the /api/challenge/item_list/ (and /api/search/) responses — the same browser-interception pattern as the Reddit and YouTube actors. There is nothing to reverse-engineer and no cookie that expires.

Features

• One keyword (query) or many (queries) — each searched separately, combined and de-duplicated by video.
• Hashtag feed (/tag/<keyword>, default, open to logged-out visitors) or full-text search mode. Whitespace collapses into a single hashtag (artificial intelligence → the #artificialintelligence feed).
• Auto-pagination by scrolling until numberOfPosts per keyword is reached.
• Actor-side timeSince/timeUntil date filtering (TikTok has no date operators).
• Anti-detection built in: stealth browser fingerprint, residential-proxy preflight, and automatic relaunch on a fresh IP when TikTok serves a captcha / empty responses.
• Results pushed per keyword, so an abort keeps everything collected so far.

Input

{
    "queries": ["econet", "#zimbabwe"],
    "numberOfPosts": 100,
    "searchMode": "hashtag",
    "region": "US",
    "proxyConfiguration": {
        "useApifyProxy": true,
        "apifyProxyGroups": ["RESIDENTIAL"],
        "apifyProxyCountry": "US"
    }
}

numberOfPosts applies per keyword. Keep the RESIDENTIAL proxy — TikTok blocks datacenter IPs and serves empty/captcha responses to flagged exits.

Output

One dataset item per video. Fields are emitted in both clean and snake_case/camelCase forms so the item is a drop-in for the powerai/tiktok-hashtag-posts-scraper output that scraping-tool's normalizeSocialPost already understands.

{
    "id": "7401234567890123456",
    "url": "https://www.tiktok.com/@zimtech/video/7401234567890123456",
    "webVideoUrl": "https://www.tiktok.com/@zimtech/video/7401234567890123456",
    "text": "AI is changing Zim business #econet",
    "author": {
        "uniqueId": "zimtech", "unique_id": "zimtech",
        "nickname": "Zim Tech", "id": "99",
        "profileUrl": "https://www.tiktok.com/@zimtech", "verified": true
    },
    "create_time": 1782898212,
    "created_at": "2026-07-01T09:30:12.000Z",
    "likesCount": 4200,  "diggCount": 4200,   "digg_count": 4200,
    "commentsCount": 88, "commentCount": 88,  "comment_count": 88,
    "sharesCount": 12,   "shareCount": 12,    "share_count": 12,
    "viewsCount": 150000,"playCount": 150000, "play_count": 150000,
    "duration": 42,
    "hashtags": ["econet", "ai"],
    "searchQuery": "econet",
    "hashtag": "econet"
}

created_at is exact (TikTok exposes real upload timestamps, unlike YouTube search).

Integration (scraping-tool)

Drop-in for the existing TikTok keyword path — this is our own actor with the same output shape as powerai/tiktok-hashtag-posts-scraper, so set it as primary with the vendor as fallback (mirroring Facebook), and normalizeSocialPost needs no change:

// ACTOR_FALLBACKS in scrapingService.js
'outspoken_strategy/tiktok-post-search-scraper': 'powerai/tiktok-hashtag-posts-scraper',

// socialProjectMonitoringService.scrapeSocialPostsByKeyword (tiktok branch)
const res = await this.scrapingService.scrape({
    url: `https://www.tiktok.com/tag/${encodeURIComponent(keyword)}`,
    resultsLimit: requestCount,
    actor: 'outspoken_strategy/tiktok-post-search-scraper',
    timeout,
    additionalInput: {
        query: keyword, numberOfPosts: requestCount, region: 'ZW',
        proxyConfiguration: { useApifyProxy: true, apifyProxyGroups: ['RESIDENTIAL'], apifyProxyCountry: 'ZW' }
    },
    scrapeType: 'tiktok-keyword-search'
});

The existing normalizeSocialPost already reads author.unique_id, video_id/id, create_time, digg_count, comment_count, share_count, play_count — all present.

Local development

npm install
echo '{ "query": "artificial intelligence", "numberOfPosts": 30, "proxyConfiguration": { "useApifyProxy": true, "apifyProxyGroups": ["RESIDENTIAL"], "apifyProxyCountry": "US" } }' > storage/key_value_stores/default/INPUT.json
npm start

Note: TikTok flags IPs that make many requests, so local runs from a home IP may be served empty responses — the actor is designed to run on Apify residential proxies. Deploy with apify push.