Restrictions are per-surface, not one switch
"Shadowban" is not a single state. In practice creators report several independent restrictions: your account missing from search suggestions, your posts missing from search results, your replies hidden from non-followers, and your replies ranked far below newer replies on other people's posts. An account can have one of these without the others.
This is why a single "banned yes/no" answer is almost useless. The useful question is: which surfaces are affected, and how strongly is the evidence?
What can actually be measured from the outside
Search suggestion: type the exact handle and see whether the account appears. This is measurable, and absence of an established account from its own exact-handle suggestion is a meaningful negative — we control-test known accounts to validate it.
Search results: post something, then search distinctive text from it. If the post never surfaces while the account clearly has recent posts, that is a measurable failure. The reverse (the post appearing) is weaker evidence, because indexes differ between regions and interfaces.
Reply visibility: open the reply thread of a post you replied to, from a logged-out view. If your reply exists but does not show up while strangers' replies do, that is a hard, reproducible signal.
What cannot be measured honestly
Reply ranking is the hard one. X orders replies by its own relevance signals — engagement, relationships, Premium status, and more. A tool outside X cannot prove your reply was demoted versus merely sorted below newer or more engaged replies. Any site claiming a precise "ranking penalty score" is guessing.
That is why an honest checker reports reply ranking as untestable rather than inventing a number. Every check on this site prints its confidence level and its limitations, and unmeasurable checks are excluded from the score instead of being faked.
Restrictions are usually reversible
Most restriction states we observe are spam-filter class, not account termination: they respond to behavior. Stopping the triggering pattern (automation bursts, duplicate content, follow/unfollow churn, link-heavy posting) and returning to normal human activity typically lifts them within days to a couple of weeks.
The practical loop: measure which surfaces are affected → remove likely triggers → re-measure after several days. Repeat once. If nothing changes after two weeks of clean behavior, appeal through official channels — and be skeptical of anyone selling a guaranteed "unban" service.