Shadowban Checker: TikTok Reach Checks on the Web
Shadowban Checker is our own product. Creators connect their TikTok account with TikTok's official login, and we read the numbers of their recent videos and return a 0 to 100 risk score with the warning signs we found. It runs as a web dashboard in 6 languages.
- Project
- Creator tool
- Our role
- Own product, end-to-end
- Platform
- Static web
- Country
- Global
- Year
- 2026
The challenge
TikTok never says
When a creator's views drop, TikTok gives no explanation. There is no official signal that an account is limited, so creators search for a "shadow ban check" and find tools that give a confident yes or no without showing how they got there.
Honest about what can be measured
A useful checker can only work backwards from results: views, likes, comments and posting patterns. It had to separate a real drop from normal ups and downs, and say plainly that its score is a reading of the numbers, not a message from TikTok.
Read-only, and easy to trust
Creators are wary of handing their account to a third party. The product needed TikTok's official login with read-only access, nothing to install on the web, and clear pages on what data is read and why.
Our approach
A score built on the creator's own baseline
The check compares the 3 newest videos with the median (middle value) of the creator's older videos. It uses the median absolute deviation, so one viral or flopped video does not skew the result. Several signals add to a single 0 to 100 score. They include a sudden fall-off, fewer views while likes and comments per view rise, and new videos that stall. Hashtag risk is shown separately and never counted as a ban.
Web dashboard
The site is plain static HTML on Vercel, with no build step. Serverless functions handle TikTok login, re-checks and billing, and Supabase stores accounts and history. The dashboard has an overview, a history view, a findings view and settings.
Pro as a watch, not a quota
Free users get 2 checks a day. Pro checks the account every day and emails within a day if views drop, with views over time and a full history. On the web, subscriptions are billed through Creem as merchant of record.
Content that explains the limits
A public methodology page lists every signal and what the score cannot tell you. A shadowban guide, a glossary, a comparison of checkers, 10 country rule guides and a blog answer the questions creators search for. The site is translated into English, Spanish, Portuguese, Japanese, Turkish and Arabic.
Screens
Inside Shadowban Checker
How we delivered
From first brief to running in production
-
01
DataScoring engine
We built a risk score that compares new videos with the creator's own baseline and lists each warning sign it found.
Median baselineMAD0–100 score -
02
Web dashboard
DevelopmentWe shipped a read-only TikTok connection, re-checks, history and daily monitoring with email alerts.
TikTok Login KitVercelSupabase -
03
Method & six languages
ContentWe published the scoring method, a shadowban guide and country rules, and generated the whole site in 6 languages.
Methodology10 country guides6 languages
Results
What is running today
Live on the web
Shadowban Checker runs on shadowbanchecker.com with TikTok's official login. Every check returns a score, a plain-language label from "normal" to "likely limited", and the signals behind it.
Six languages from one build
The site's pages and blog are generated in 6 languages from one English source. They have hreflang links (tags that point search engines to each language) and an Arabic right-to-left layout. Once a visitor picks a language, the menu never switches it on its own.
A transparent method
Instead of claiming certainty, the product publishes how the score works, what data it reads and what it cannot know. A quick in-browser self-check on the homepage lets visitors try it before connecting anything.
More work
Other things we built and run
SchoolVita
SchoolVita is our own product: a directory of private and international schools in 20 cities where families relocate. Parents compare estimated fees and curricula, then send an enquiry to the schools they choose. Schools can claim their listing and receive those enquiries.
Senfoni Koçluk
An Istanbul student-coaching practice ran on WhatsApp, spreadsheets and memory. We built one Laravel platform that covers the public site and free exam tools, portals for students and parents, and the coach's back office: exams, schedules, sessions and payments.
PlantCareClub
PlantCareClub is our own product: a houseplant care site with plant profiles, step-by-step guides, free diagnostic tools and a curated gear shop. Visitors can save their plants to a free My Garden account. We build and run all of it on Laravel.
