You can test for a Twitter shadowban in about two minutes, from a logged-out browser, without any third-party checker. Search from:yourusername. If your recent posts do not appear there but your followers still see them in their timeline, you are not imagining it: that specific combination is what X calls being filtered from search.
The word "shadowban" is not X's word. But the behaviour it describes is documented, and X states the conditions that trigger it.
How to test for a Twitter shadowban yourself
Do this before changing anything about your account. Three tests, all free, all from your own phone.
1. The search test
Open a private browsing window, or use an account that does not follow you. Search:
from:yourusername
If your recent posts do not appear, your account may be filtered from search. If they appear normally, search visibility is fine and the problem is elsewhere.
2. The exact-phrase test
Take a distinctive phrase from a recent post, something unlikely to appear anywhere else, and search it in quotes from that same logged-out window. If the post does not come up, the filtering is at the post level rather than the account level.
3. The reply test
Reply to a popular post from your account. Then open that thread from a different account, or logged out. If your reply is collapsed, hidden behind "show more replies", or absent, your reply visibility is affected.
The signature that confirms it
Here is the detail that separates a real filtering event from a bad week, and it comes straight from X.
The search policy states that when an account is filtered from search, your followers will still receive your updatesOpens in a new tab and direct mentions are still delivered. X's help page on restricted accounts is even more explicit about limited accounts: "only your followers are able see your activity on X, including posts, likes, reposts, etc."
So the signature is asymmetric. Engagement from existing followers looks roughly normal. Everything from outside your follower graph collapses. If both dropped together, you are probably looking at a content problem, not a filtering problem.
What X actually calls it
The platform does not use the word shadowban, which is part of why the topic is so confused. X talks about search filtering, limited accounts, platform manipulation and enforcement.
That matters practically: if you go looking for a "shadowban" setting or notification, you will not find one. What exists is a set of documented conditions under which visibility is reduced.
Why accounts get filtered from search
Duplicate posts and duplicate links
This is the one most people trip over, and X states it plainly. The search policy says accounts may be filtered for posting similar messages over several accountsOpens in a new tab, and that repeatedly posting near-duplicate updates or duplicate links can get you filtered.
Read that twice if you run more than one account. The mechanism is not judging whether your offer is good. It is detecting that the same message and the same URL appear in several places, which is a pattern, and the pattern is what gets filtered.
Behaviour that reads as automated
Aggressive follow and unfollow cycles, mass replies, high-volume posting on a fixed schedule, identical calls to action. None of these are individually fatal. Together they describe a bot, and the system is not weighing your intentions.
Reports, blocks and mutes
These accumulate quietly. You do not get told. A run of blocks after a promotional push is a signal, and it is one of the few that you can influence only by changing what you post.
Is it a shadowban, or is it your links?
This is the test almost nobody runs, and it is the most useful one in the article.
Compare your link posts against your text posts
Post the same idea twice over a few days: once as plain text, once with a link. Then compare impressions.
If both perform badly, the account has a visibility problem. If your text posts are fine and your link posts die, your account is not filtered. Your links are the issue. Those are two completely different problems with two completely different fixes, and treating one as the other wastes weeks.
Why the same URL everywhere is a liability
Two separate mechanisms stack here.
The first is the one X documents: duplicate links across accounts are a search-filtering trigger, by their own policy.
The second is domain reputation, which is not specific to X at all. A domain shared by many accounts carries the consequences of all of them. One report against any of those accounts makes the address suspect, and every account pointing at it inherits the problem. We take that apart in why a custom domain matters for bio links.
The practical rule that follows is dull and effective: one domain per account. Not one per brand, not one per campaign. It costs a few dollars a year and it turns a shared failure into an isolated one.
How to recover
There is no button, and anyone selling you one is selling hope with a checkout page.
Stop first, then act
Pause the duplicate posting. Stop posting the same URL repeatedly. Turn off automation for a few days. Cut the mass replies. This is not a fix in itself, it stops the signal from accumulating while you work on the rest.
Then give the account normal behaviour
Post original content without links for several days. Reply like a person. No copy-paste, no promotional pressure. You are rebuilding a trust signal, and trust signals rebuild on time and consistency, not on intensity.
Remove recent posts that are clearly duplicated or spam-like. Do not delete your whole history in a panic. A sudden mass deletion is itself an unusual pattern.
Retest, do not guess
Run the three tests from the top of this article again after a few days. That is the only feedback loop you have, so use it instead of watching the impression counter.
If the account is locked or limited
That is a different state from search filtering, and it has an official route. X's help page on locked or restricted accountsOpens in a new tab covers what each state means and how to appeal if you think the decision is wrong. Use it. It is the only channel that actually reaches a decision-maker.
What does not work
Third-party shadowban checkers. Some give a usable hint, but none of them can see X's internal state. They infer it from the same public searches you can run yourself, and they are wrong often enough that a false positive will send you rewriting an account that was fine.
Deleting everything. Covered above, but it comes up constantly.
Buying engagement to "reset" the algorithm. This adds low-quality engagement signals to an account that is already being judged on signal quality.
Making a new account and pointing it at the same link. This is the fastest way to turn one filtered account into two, because the duplicate-link pattern is exactly what X says it filters for.
Where the link layer fits
If your account earns from sending people somewhere, the link is not a detail of the account, it is part of it. That means a domain you control rather than one shared with strangers, one domain per account rather than one for all of them, and enough measurement to know whether a click actually arrived.
That is the part LinkScale handles, and the reasoning behind the routing side is in what is a deep link.
Frequently asked questions
How long does a Twitter shadowban last?
There is no published duration, because there is no formal state called a shadowban. Search filtering lifts when the behaviour that triggered it stops and enough time passes, which in practice means days rather than hours. A locked or limited account is different and follows the appeal process instead.
Can a link get my account filtered?
Duplicate links are named in X's search policy as a filtering trigger, so yes, repeatedly posting the same URL can contribute. A single link in your bio is not the problem. The same link posted across several accounts is.
Is a shadowban the same as a suspension?
No. Search filtering leaves the account fully usable and visible to followers. A suspension removes access. A limited account sits between the two, with X stating that only followers can see your activity.
Should I delete my tweets?
Remove recent posts that are duplicated or clearly spam-like. Leave the rest. A mass deletion is an unusual pattern in itself and removes the history that makes an account look real.
Do shadowban checker sites work?
They run the same public searches you can run yourself and infer the rest. Use them as a hint at most, and run the three manual tests before you act on anything.
Why do I still get likes from followers but no new reach?
That asymmetry is the documented signature of search filtering. X states that filtered accounts still reach their followers, which is exactly why the drop looks like the algorithm suddenly hating you rather than a visibility restriction.
Last updated: 5 September 2026. Written by Thomas Melto, CEO of LinkScale.
