X Under the Hood vs shadowban checkers

X built Under the Hood as a transparency report about labels, not as a button that says 'you are shadowbanned' or 'you are not.' You open it from the account you care about, at x.com/i/under_the_hood. When X expanded the report in September 2026, the company described it as a way to see labels applied to your account and posts in the prior month, plus cases where visibility was limited because X said it was complying with local law, including which country was involved in a legal demand.

The report is an aggregate. For post labels, you see that a label existed and how it showed up across the period, not a feed of the status IDs that carried it. Account labels are summarized the same way. X has also offered the underlying data as a JSON download for people who want the raw file. A non-technical reading of the on-screen report is enough for the comparison this page is about.

Labels are the same family of controls X talks about when it describes restricted reach. A label can mean a post was downranked, kept out of recommendations to people who do not follow you, or otherwise shown to fewer people. The report is where X names those labels for your account. It is the closest thing to an official record. It still does not use the vocabulary of search bans and ghost bans that checkers invented, and a month with no labels is not a certificate that every reply ranked well.

Government-required limits are a separate section of the same idea. If X withheld a post in a country or reduced visibility because of a legal demand, the expanded report is meant to show that, including the country. That is not the same event as a spam-score on your replies. Read the reason the report actually gives you instead of folding every row into the word shadowban.

When X widened access in September 2026, it said the report was available to accounts that meet a basic bar: at least 10 posts in the prior month, and an account that is at least one year old. If the page does not offer the report, assume you are outside that bar or that X has changed the rule since. The page itself is the source of truth. This guide cannot see your eligibility.

That bar leaves out a lot of people who are the most worried: new accounts, and accounts that post rarely. Those accounts can still have a bad week in search or in replies. Under the Hood will not explain it. They need the logged-out tests in the manual shadowban test, not a refresh of a report they cannot open.

You can only pull the report for the account you are logged into. There is no public Under the Hood URL for someone else's handle. If a tool claims it downloaded X's official label file for an arbitrary username, it is not describing this feature. Anything it shows you is an external guess.

Suspended accounts and accounts that have not posted enough in the window can also fail the post-count requirement even if the account is old. If you cannot log in normally, you are dealing with a suspension or a lock, which X surfaces directly. Under the Hood is not the appeal form for that.

Start with the period. The report looks backward over a month, not at the reply you sent after lunch. A label count of zero for that period means X is not listing labels in the aggregate. It does not mean this afternoon's post has finished indexing, and it does not mean a logged-out from: search will succeed for every post.

When labels are listed, read the name, the description, and the effect on visibility as X wrote them. Those sentences are more specific than 'shadowban.' A spam-oriented label that keeps posts out of recommendations to non-followers is a distribution limit. It may line up with what you see in replies or search, and it may not line up with every checker badge. Quote the label name in your notes. Do not translate it into a different product's brand of ban unless the effect text actually matches the test you ran.

The report does not identify which posts made up the count. You cannot jump from 'twelve posts carried a label' to a particular URL. If you need the URL, you still have to look at recent posts yourself for warnings, and you still have to run search and reply tests. The sensitive-media probe on a checker is sometimes more specific about recent posts than Under the Hood is, and less authoritative, because it is inferring from a public field rather than reading X's label store. Both limits are real. They are described in the sensitive media guide.

A checker such as the one on the homepage does not download your Under the Hood file. It behaves like a stranger with a search box. It asks whether posts show up for from:handle, whether the handle is suggested, whether replies look hidden or buried, and whether recent posts carry a sensitive flag. You can run it on any public handle. You get an answer in one sitting, and the answer can be wrong when a request fails or when ranking is ordinary.

The table is the whole disagreement people have online. One person says 'Under the Hood shows no labels, so the checker is a scam.' Another says 'the checker shows a search miss, so X is lying.' Both tools can be behaving as designed. No labels in a monthly file does not replay today's Latest tab. A checker fail does not prove X stored a label. When they disagree, do the manual test logged out. That test is the same family as the checker, done slowly enough that you can see Show more and Latest with your own eyes.

Checkers also disagree with each other. One never tests replies. One calls every Show more placement a ghost ban. One caches a result for many minutes. Under the Hood will not referee that fight. It does not implement their categories. Map each badge back to a concrete test, using the search, ghost, and reply deboost pages, before you average three logos into a verdict.

1. If you can open Under the Hood, read it once and save the label names and effects. Note the month it covers.
2. Run the logged-out search, autocomplete, and reply tests, or run the checker and then confirm any fail by hand.
3. Line them up. A label whose effect mentions recommendations or ranking supports a reach problem you can also see in public. A public search miss with no label is still a public search miss. A label with no public miss is a reason to read the effect carefully, not a reason to panic-delete the account.
4. Change the behavior that matches the public symptom. Retest the public symptom later. Look at a later month's report if you want to see whether X is still listing the label. Do not expect the report to update because you pressed Check.

If you are not eligible, skip step one instead of inventing a workaround that scrapes someone else's report. You still have the manual test and the checker. Eligibility is X's gate, not a hint that quiet accounts are exempt from ranking.

Recovery steps stay the same either way: verify the account, stop burst behavior, remove specific flagged posts and spammy links, and retest. The sequence is in how to recover. Under the Hood does not add a hidden fifth step that support will execute for you.

Under the Hood will not show you another user, will not update live, will not list post URLs, and will not open for every account. It can omit detail X is unwilling to expose because that detail would help people evade enforcement. A sparse report is not a bug in your browser until the page itself says the report is unavailable.

A checker will not show you X's label name, will not explain a legal withhold, and will not be consistent across vendors. It can be run on a handle you do not own, which is also its privacy boundary: it should only be reading what a stranger could see. This site's result pages at /check/handle are noindex bookmarks of a heuristic, not official reports.

Use Under the Hood when you are eligible and you want X's wording. Use the checker when you want a fast pass over public symptoms for any handle. Use the logged-out tests when you need to settle a disagreement. None of the three tells you how many days a limit will last.