Pull TikTok data
like the algorithm does.
Extract structured data from hashtags, creator profiles, search queries, and individual video URLs — views, likes, shares, comments, music, followers, downloads. Pick targets, queue a job, get a clean dataset back in under a minute.
From targets to dataset in three steps.
Pick your targets
Hashtags (#fyp, #booktok), creator handles (@username), free-text search, music IDs, or paste TikTok URLs directly. One job can mix targets — five hashtags across the same query are fine.
We run the scrape
Job lands in a queue. We hit TikTok's public surface via our distributed scrape network — no rate-limit walls, no captcha headaches. Progress is live; you can close the tab and we'll email or webhook you when it's done.
Export anywhere
Download as CSV, JSON, Excel, or HTML, fire a webhook into your warehouse, or hit the REST / MCP / x402 API. The dataset is yours — no DRM, no usage cap on what you've already paid for.
Thirty-plus fields per post. Clean. Normalized. Queryable.
Every row is the same shape no matter what you searched for. Author stats and creator profile come back in one object. Music metadata sits in its own. Hashtags, mentions, and effects arrive as arrays. Subtitles and video URLs are pre-resolved so you can download them in a follow-up step without a second scrape.
From "catch this trend by Friday" to "this is our weekly data feed."
One scraper, six workflows, one queryable dataset. Most teams start with one and end up running four of these at once.
Trend hunting
Catch a trend the morning it pops. Pull the top posts under #booktok, #cleantok, or any niche tag and rank by velocity (plays per hour). Schedule weekly re-scrapes so trends never sneak past you.
Creator research
Build qualified creator lists. Pull every post from a profile, sort by engagement rate, calculate average plays per post. Add follower lookups to see who they reach and who they follow back.
Brand monitoring
Track every mention of your brand, hashtag, or product name. Watch for unsolicited UGC, catch counterfeits in real time, find micro-influencers already talking about you.
Music discovery
Pull every video using a specific sound. Sort by play count to find the breakouts. Pull the music ID and reverse-lookup the original creator. Useful for labels and supervisors hunting cues.
Geo-aware research
Run the same search from a US, UK, JP, or BR proxy and see how the algorithm shifts. Map regional trends, surface culturally-relevant variants, find country-specific creators in seconds.
Dataset operations
Bulk-load posts into your warehouse. Webhook → Snowflake / BigQuery. The same job spec via REST or MCP, deduped and idempotent. Build LLM training sets, recommendation models, or trend forecasts.
REST, MCP, webhook, or pay-per-call x402.
Ounie accounts get a REST API with bearer-token auth and a native MCP server that drops into Cursor, Claude, ChatGPT, or the AI SDK — your assistant can scrape hashtags, poll a job, and export results without writing any glue. One-off scripts hit our x402 endpoint and pay per request in USDC.
// One request. Streaming results.
const res = await fetch("https://tiktok-scout.ounie.com/api/scrape", {
method: "POST",
headers: {
"Authorization": "Bearer ${API_KEY}",
"Content-Type": "application/json",
},
body: JSON.stringify({
mode: "hashtag",
targets: ["fyp", "booktok"],
max_items: 1000,
download_comments: true,
}),
});
const { job_id } = await res.json();
// → poll /api/jobs/${job_id} or wait on the webhookPay for results. Not seats, not setup, not surprises.
One scraped result costs 2 Ounie credits (+1 with comments, +1 with follower lookup) — the same wallet every Ounie product shares. You are only charged for results actually returned; unused reserves refund automatically. AI agents can skip the account entirely and pay per call in USDC via x402.
The questions we get on the first call.
Sign in with Ounie, run your first scrape in 90 seconds.
Pay for results in credits, not seats. No subscription — keep every row you've already pulled.