X shadowban: how to test for one, and what actually fixes it

If your replies stopped getting answers and your handle stopped appearing when people type it, you are probably not imagining it. X limits reach as a deliberate enforcement action. It does not call the result a shadowban, but the list of what it does is public, and so is the test.

This is the explanation the checker tools do not give: what X (formerly Twitter) actually does when it limits reach, the three things people mean by shadowban, the manual test that works when the tools are down, and what changes it. It does not promise recovery.

What a shadowban is on X in 2026

X never uses the word. It publishes a list of enforcement options, one of which describes the thing exactly. The enforcement options page (checked 19 September 2026, as were all the help pages quoted here) puts it under limiting post visibility: "Where appropriate, we will restrict the reach of posts that violate our policies and create a negative experience for other users by making the post less discoverable on X. This can include:"

• Excluding the post from search results, trends, and recommended notifications
• Removing the post from the For you and Following timelines
• Restricting the post's discoverability to the author's profile
• Downranking the post in replies
• Restricting Likes, replies, Reposts, Quote posts, bookmarks, share, pin to profile, or Edit post
• Excluding the post from having ads adjacent to it

Everything people report as a shadowban is on that list. The unit is the post, and the page adds: "Starting in April 2023, posts identified as violating our Rules will begin to receive labels informing both post authors and viewers that we limited the post's visibility." Also: "Authors will be able to submit an appeal on the label". A limit on one post is visible and appealable; the account-wide version operators run into carries no label and no appeal button, which is why the word survives.

The three things people mean

Search suggestion ban

The handle no longer appears in the suggestions that drop down as someone types it; the profile still loads directly. It is the mildest, the most common, and what most checker tools really test.

Search ban

A search for from:handle returns nothing, or nothing recent, from a session that is not the account itself. The posts exist; search will not surface them. With the first, this is what most people mean by shadowbanned.

Reply deboosting

The reply shows on the account's own profile, but in the thread it sits below the fold, behind "Show more replies", or is not shown to others at all. People call these ghost replies. It is "Downranking the post in replies" applied to an account, and it hurts reply-driven accounts most: the account keeps working and simply gets no answers.

Two things that are not shadowbans

Read-only mode, from the enforcement page: "We may temporarily limit an account's ability to post, Repost, or Like. The user can read their timelines and will only be able to send Direct Messages to their followers." If you cannot post, it is this.

A lock is the other. The locked and limited accounts page says visitors "may see a message letting them know the account has displayed unusual activity and asking them to confirm they still want to view it", and the enforcement page adds that a locked account "is removed from follower counts, Reposts, and Likes". It looks like a punishment for reach; it is a verification challenge, best cleared without making it worse: unlocking a locked X account.

How to test without a tool

The whole test rests on one rule: the session doing the looking must not be the account being looked at. Your own session always sees your own content, regardless of any limit.

Open a private window with no X login, or log in to a second account that does not follow the one you are testing. If a logged-out window is refused search results, use the second account. Then, in order:

1. Suggestions. On the explore page, type the handle into the search box, character by character. A healthy account appears in the drop-down before you finish; if it never appears and a comparable account does, that is a suggestion ban. Compare, because a brand new account with no followers can be absent for other reasons.
2. Search. Search from:handle and switch to Latest. Recent posts should be listed. If not, run the same search logged in as the account. Posts visible inside and not outside are the definition of a search ban.
3. Replies. Open a post the account recently replied to, from the second account. Look in the thread, then under "Show more replies". A reply visible only to the account that wrote it is being deboosted. Check two or three threads; one reply can sit low simply because nobody engaged with it.
4. Profile. Open the profile. A warning about unusual activity means locked, not shadowbanned, and the route out is verification. A suspended banner means this page does not apply.

flottra runs the first two checks on the accounts it operates, on a schedule, and moves an account that fails either into a cooling state with reduced activity until it passes again. It checks only its own accounts; it cannot look at a handle it does not run.

Why the checker tools disagree with each other

Search for a shadowban test and the results are checker sites: shadowban.yuzurisa.com, HiSubway, Circleboom and others. They do a version of what you just did, from a session they control, and they disagree, with each other and with themselves a minute later, for three reasons. That makes them a second opinion, not a verdict.

• They scrape the same two searches. Suggestions and from:. There is no shadowban API; a checker is a script running those searches and parsing the page.
• X changes the interface. When the search page changes the scrapers break, and a scraper that finds no posts often reports a ban rather than an error.
• Search results are personalised. What X returns depends on the session asking: login state, history, location. The tool's session is not your reader's, so a result honest for the tool can still be wrong for you.

What causes it when you run several accounts

X publishes the enforcement list, not the thresholds. What follows is observable from watching multi-account setups get limited: pattern, not rule.

• The same text from several accounts. Duplicate and near duplicate posts and replies link accounts and get all of them limited at once. The authenticity policy calls it content spam: "You may not share or post content in a bulk, duplicative, irrelevant or unsolicited manner that disrupts people's experience."
• Reply volume. Replies are where limits bite first; downranking in replies is a named enforcement option. An account that replies far more than it posts, to people who never engaged with it, is the shape the automation rules describe: "sending automated replies to posts based on keyword searches alone is not permitted". The reply guide covers what a reply engine needs to survive.
• Links in replies. A reply with a link, from an account whose replies mostly carry links, is the shape of spam whether or not the link is.
• Engagement spam. Likes, reposts and follows exchanged between your own accounts. The same policy: "We prohibit inauthentic use of X engagement features to artificially impact traffic or disrupt people's experience."
• Reports. Enough people reporting your replies is a signal you cannot pace away, and crowded topics get reported more.
• New accounts with high output. Nothing, then a burst of everything. This is why the warming ramp in the suspensions guide starts with three days of reading and no posts.

What actually fixes it

A shadowban is lifted by X, on a schedule it does not publish, after the behaviour that caused it stops. Nobody outside X knows the duration. You control the behaviour, and whether anything else is pending on the account.

1. Stop the behaviour, not the account. Find which pattern above the account was running and stop that one. Going silent is not required; posting like an ordinary person is.
2. Slow down for days, not hours. Halving output for an afternoon changes nothing X would notice. Days are the unit.
3. Original text per account, replies included.
4. No links in replies for a while. Links go in posts on the account's own profile; replies stay plain text until reach returns.
5. Complete anything pending. Log in as the account, from its usual proxy and browser profile, and look for a verification prompt or a label. A label can be appealed; a pending verification is a lock, and nothing else moves until it is cleared.
6. Re-test from a second account once a day. Not hourly. The test does not change the outcome.

Across many accounts

A shadowban on one account is information about the others. Accounts that share text, a proxy or a target list tend to get limited together. Test the siblings before you assume.

The check also belongs on a schedule, because reach loss is silent by design. Cooling a failing account, as flottra does, does not lift the limit or promise the account comes back; it removes the week of posting into a void before anyone noticed.

The rest is the list that keeps accounts from being suspended, because a shadowban is often the warning before one: one proxy per account, a persistent browser profile, caps well under the visible limit (the rate limits guide puts the official numbers next to what a cautious operator runs), and text nobody else is posting.