How to Check if You Are Shadowbanned on X (and What Each Result Means)

You Cannot Check This From Inside Your Own Account

This is the part almost every "am I shadowbanned" guide skips, and it makes most of the advice useless.

X personalises what you see. Search your own handle while logged in and you will find yourself at position one, usually with "You follow each other" underneath. Open one of your own replies and it is right where you left it. None of that tells you anything, because you are the one person X will always show your own content to.

The only way to answer the question is to look from outside: from accounts that do not follow you, have never interacted with you, and have no reason to be shown your posts. That is what Shadowban Check does. It is free, there is no login, and nothing about the check is stored.

The Five Things It Looks At

X does not have one switch called "shadowban". It has several different limits that people lump under that word, and they fail independently, so an account can be invisible in search while its replies are perfectly fine.

1. Search suggestion

When someone starts typing a name into X's search box, X offers a dropdown of accounts. The check asks X for that list for your exact handle and looks for you in it.

The only question is whether you come back. People sometimes argue that a missing account just lost a ranking contest to bigger ones with similar names, but that is not what the data looks like. On one account we measured, X returned nine near-misses of the handle, variations with digits and underscores, and not the exact match. A ranking system does not drop the exact string while keeping the typos of it.

2. Search ban

A search for your posts returns nothing, while you have actually been posting.

That second half matters more than it sounds. An account that has not posted in months has nothing for search to return, and calling that a ban is the easiest false positive there is. The check only runs this row if you have posted in the last 45 days, and says so plainly otherwise.

3. Ghost ban

Your replies inside other people's conversations stop being visible to anyone who does not follow you. The thread still exists, your reply still exists in your own timeline, and nobody else in that conversation can see it.

The check takes replies you wrote in other people's threads, opens those conversations the way a stranger would, and looks for them. Replies inside your own threads are skipped deliberately: those are always visible and would hand back a free pass.

There is a second way the same ban shows up, from the other end. Someone replies to your post, and your post no longer appears above their reply as the thing they were answering. The thread is torn apart from both sides. Some tools call that a thread ban; it is the same limit, so it feeds the same row here rather than inflating the count with a sixth check.

4. Reply deboosting

Your reply is not missing, it is filed. X has a "Show probable spam" link at the bottom of busy conversations, and a deboosted reply sits behind it, where almost nobody clicks.

This is worth separating from a ghost ban because the fix is different and the severity is different. The check follows that link and looks for your reply there, two levels deep, because X puts replies from accounts it classes as offensive behind a second barrier inside the first.

5. Reach

Not a ban, a reading. The median views of your recent original posts, that figure as a share of your followers, and two comparisons: against other accounts of your size, and against your own earlier posts.

The first comparison is the one most tools skip, and it is the one that answers the question people actually ask. "331 median views" means nothing by itself, and neither does "0.1% of your followers" until you know what everyone else gets. We measured a sample of real accounts the same way the check measures yours, grouped by follower count, because the share falls steeply as an account grows: a small account reaching 7% of its followers is ordinary, while a 300k-follower account reaching 3% is doing fine. If your share sits below nine in ten accounts your size, the row says low reach even when your week-on-week numbers are flat, because a flat line at the bottom is still the bottom.

Two things keep this honest. Posts under 24 hours old are excluded, because they have not finished earning their views and including them makes every healthy account look like it is falling. And accounts whose posts get single-digit views get no reading at all, since a swing of one or two readers would move the whole number. Below 100 followers there is no share-of-followers reading either: nine views on three followers is 300%, which is arithmetic rather than reach.

"Could Not Check" Is Not a Pass

Every row can come back as "Could not check", and the tool deliberately does not round that up to good news.

Two things cause it. Either no reader would serve that particular request, which says nothing about your account, or there was genuinely nothing to look at. The second one is common on the reply rows: if you only ever post and never reply inside other people's conversations, there is no reply for anyone to hide, and the only honest answer is that the question does not apply to you.

A tool that turns "I do not know" into "you are fine" is worse than no tool, because you act on it.

Why Two Checks Can Disagree

X does not answer identically to every session, and this surprises people enough that it is worth saying with numbers.

On one search-banned account we measured on 19 October 2026, 1 reader in 24 still returned its posts while the other 23 saw nothing. A healthy control account returned posts 8 times out of 8. So a single reader finding your posts is not proof that search works for you; it is one session out of two dozen whose settings differ, most likely with the sensitive-content filter off.

That is why the search row takes a majority of three reads rather than trusting one, and why it tells you when the vote was split. If two checks a few minutes apart still disagree, the honest reading is that your account is sitting near the line rather than clearly on one side of it.

What No Tool Can Tell You

A check can show that a limit is there. It can never prove one is absent.

X does not publish these rules, does not notify you, and changes them without saying so. A quiet week looks exactly like a quiet account. Anyone selling you a green tick that means "definitely not shadowbanned" is selling you a feeling, not a measurement.

What a good check gives you is narrower and more useful: evidence about specific, testable things. Your handle is missing from X's own suggestion list. Your posts do not come back in search while you are posting daily. Your replies are not in the conversations you wrote them in. Those are facts you can act on, and they are the ones worth collecting before you start rewriting your bio or deleting posts on a hunch.

Try It

Shadowban Check takes a username and about fifteen seconds. Free, no account, two checks a minute so the reader pool stays usable for everybody.

If you need the same kind of live X data in your own code rather than on a page, that is what the XCROP API is for: profiles, timelines, search and interaction checks, from the same engine this tool runs on.