How FSB checks a shadowban (and avoids false alarms)

A shadowban checker is only useful if you can trust a "detected". A false positive sends people into a panic and makes them change things that were fine. Here is how FSB is built to avoid that.

Four independent tests

FSB tests search visibility, search suggestions, reply visibility (ghost ban) and reply ranking (deboosting) separately, all from a logged-out view, which is how search engines and people who do not follow you see an account. Each test has its own result, so you see exactly what is affected.

Three answers, not two

Every test returns detected, clear or unknown. Unknown happens when X does not answer, rate-limits, or returns something incomplete. Many tools treat "no answer" as "banned". FSB keeps the last confirmed state instead and tries again later.

Confirmation before publishing

A new restriction is re-tested before it is shown, so one bad response cannot flip an account to "shadowbanned". Restrictions that disappear are confirmed the same way before being cleared.

Sensitive content is separated

X hides sensitive media from some searchers. That can look exactly like a search ban. FSB reports it as the sensitive flag instead, so accounts with an old "mark media as sensitive" setting are not told they are search banned.

Several sources

Profiles, timelines and engagement counts come from X's public web data, cross-checked against open embed services such as FixTweet and vxTwitter. Official counts win when sources disagree, and partial data never overwrites complete data.

Watching over time

Accounts on the global board and in dashboards are re-checked automatically, and every result is stored, so the history graph shows when a restriction started and ended. You can see the system's own health on the status page.