Douyin
Live platform data · China · United States · Georgia
See what's actually working, before you make the next thing.
Six platforms, real counts, one argued report. You correct the brief before I spend anything.
Three reports free, no card. I read your brief back to you before I spend anything.
You write
What are people actually buying for a home coffee setup?
I read it back
Topic: home coffee gear, US market. Four search angles across TikTok, Instagram and YouTube. Report written in English.
- home coffee setup
- espresso under $500
- coffee gear haul
- morning coffee routine
That's right, run itNot quite — let me adjust
- Douyin
- Xiaohongshu
- TikTok
- YouTube
- views
- likes
- comments
- shares
- saves
- hashtag supply
- creator profiles
- comment census
- search demand
The harvest
Six platforms in. One report out.
Searched in parallel, filtered for the market you picked, ranked, and merged into a single document with its working shown.
- platforms
- 6
- markets
- 3
- report
- 1
Who it's for
For the people who have to answer “so what should we make next?”
Freelance strategists
Bill for the answer, not the afternoon.
Brand and DTC leads
Decide next month’s drop on evidence.
Cross-border operators
Read China and the West in one report.
Creator-agency scouts
Find who is working before they get expensive.
The numbers are the platforms’. The argument is mine. The working is yours.
Every report ends with its own methodology page
Sixty seconds of the actual product
Every screen in this film is a real screenshot of the running app — the brief being typed, the interpretation you get to correct, and the finished report. Nothing is a mockup.
60 seconds, no narration. Sound on for the design, but it reads fine muted.
The round trip
How it works
01
You describe it in plain language
A niche, a product category, a thing you keep seeing on your feed. No keyword syntax, no boolean filters, no platform-by-platform setup.
02
I read the brief back to you
Before anything is spent I show you the topic I detected, the search angles I'd use on each platform, and the language I'd write in. Edit any of it directly, or tell me what I got wrong and have me read it again.
03
I harvest, then I write
A couple of minutes of live searching, then a ranked and annotated read: what is working, which themes it falls into, and what I'd do about it. Every keyword searched and every call billed is listed at the bottom.
Why the second step exists
Most of what goes wrong in automated research goes wrong at the brief. The tool searches confidently for the wrong thing, and you only find out at the end, after the budget is gone. Correcting the brief costs nothing. Correcting the report costs a report.
Public reports
A finished report, in public
These came out of the same pipeline you'd use, and they're published in full — every number, every caveat, every keyword searched. No sign-in.
No public reports are published yet. When one is, it will be linked here in full rather than described.
Method
Why the numbers are real
via TikHub · 6 platforms
Every metric comes from a platform endpoint
- China
- DouyinXiaohongshuWeibo
- United States
- TikTokInstagramYouTube
- Georgia
- TikTokInstagramYouTube
The model writes the prose; the counts are computed in code
never LLM arithmetic
Thin data is reported as thin, with the item count
per platform section
itemised, in dollars
Every billed data call is logged against the report
- Keywords searched, per platform
- Every workflow step, in the order it ran
- Data calls billed to this report
- Data quality, graded on every section
comment census
Every harvested comment is classified one by one, and the distribution is counted in code
Shape only — the real distribution differs per report
Every report ends with its own methodology page: the keywords searched per platform, the work log step by step, and the number of data calls billed. It isn't an appendix — it's how you audit the answer.
Two ways to use it
The free tier is a real scan, not a teaser — it just looks at less. Pro widens the harvest and reads the comment section.
Scan
Free
3 reports, one time
For seeing whether the read is useful to you at all.
- 3 reports total, no card
- One search pass per platform, top keywords only
- Ranked posts with views, likes and engagement rate
- Themes, an opinionated take, and a methodology page
- Public share link you can send to a client
Deep scan
Most useful$29per month
20 deep reports every month
For deciding what to actually make, and for charging someone for the answer.
- 20 deep reports every month
- Every keyword your brief produces is searched — the free tier stops at the first two
- Paginated harvest across the hot board and search — three pages of the board and four of search, per keyword
- Views and comments, shares and saves joined from two endpoints that each carry only half — the report says how many posts carry the full set
- Comment census on every platform — up to 240 comments on each top post where the platform pages its comments (20 on Xiaohongshu and YouTube, which serve a single page), every one classified, the distribution counted in code rather than eyeballed
- Search-demand curve per keyword on Douyin: seven days of movement plus the adjacent terms that are rising
- Hashtag supply for your leading keywords — reach, and distinct-creator counts where the platform publishes them — so you can see how contested a tag already is
- Creator profiles behind the posts that worked — follower count, output, region — on every platform
- A per-report cost ledger in dollars, itemised by signal
What actually differs
| Signal | Scan | Deep scan |
|---|---|---|
| Deep signals available on | All six platforms, standard scan | All six — see the per-signal note |
| Keywords searched | First 2 | All of them |
| Pages per source | 2 | 3–4 per source |
| Metrics per post | Views, likes | Views, likes, comments, shares, saves (joined subset) |
| Comments read | None | Up to 240 per top post; 20 on XHS and YouTube |
| Search-demand curves | — | Per keyword, 7 days (Douyin) |
| Hashtag supply | Tag counts only | Reach per tag; creator counts where reported (4 of 6) |
| Creator profiles | — | Up to 45 per platform |
| Cost ledger | — | Itemised, in dollars |
What Pro does not do
It does not invent data. If a platform returns thin results for your topic, the deep report says so on the section, with the number of posts it actually read — exactly as the free one does. Platforms also differ in what their APIs expose: the paginated harvest, the comment census and the creator profiles run on all six, hashtag supply on four, and the seven-day search-demand curve only on Douyin, whose hot-word board is a platform feature no one else has. Every report states that coverage per platform up front, so an absent signal is never mistaken for a finding. Depth buys more evidence, not more confidence.
Start with a free scan
Three reports, no card, and I'll show you how I read your brief before I spend a single call.
FAQ
Questions
- Where does the data come from?
- Live platform data via the TikHub API — real view counts, likes and comments, never estimates.
- What languages are supported?
- Ask in English, Chinese or Georgian — the report is written natively in your language.
- How long does a report take?
- Usually 1–4 minutes. You watch every step live.
- What does the free tier actually include?
- Three reports, one time, no card. Each is a real scan: the leading keywords per platform, ranked posts with views and likes, themes, an opinionated take, a methodology page and a public share link. It isn't a preview with the numbers blurred out.
- What does Deep scan add?
- Every keyword your brief produces gets searched instead of the first two, every source is paginated, and the comments on the top posts are read and classified — on all six platforms. $29 a month for 20 deep reports. Depth varies by what each platform's API exposes: up to 240 comments per post where comments page, 20 where they don't; hashtag supply on four of six; the seven-day search-demand curve only on Douyin. Every report states that coverage per platform before it shows you anything.
- Which markets can you research?
- Three: China (Douyin, Xiaohongshu, Weibo), the United States and Georgia (TikTok, Instagram, YouTube). The report is written in the language you wrote your brief in, whichever market you picked.
- What if there isn't much data on my topic?
- The report says so, on the section, with the number of posts it actually read. It doesn't fill the gap with the model's own guesses. That holds on both tiers.