Are you shadowbanned on X (Twitter)? How to check

In 2018 Twitter gave "the best definition we found": "deliberately making someone's content undiscoverable to everyone except the person who posted it, unbeknownst to the original poster." Then it said: "We do not shadow ban." X's help pages still avoid the word, but they say X limits reach: "We may limit a post's reach if it violates our X Rules or applicable laws, or is determined to be low quality." They add: "We are constantly running experiments on X which may limit the reach of certain posts."

X's spam policy lists what that can mean: "excluding posts from search results, trends, and recommended notifications, removing posts from the For You and Following timelines, restricting a post's discoverability to the author's profile… and downranking the post in replies." What people call a shadowban is any of those, without being told.

X started piloting the report on 13 August 2026 and opened it on 18 September 2026 to all accounts "that meet the basic eligibility criteria (10+ posts in prior month, account at least one year old)." Open x.com/i/jf/under_the_hood while logged in. It lists the labels that can limit visibility on your account and your posts in the previous month, with how many of your posts got each, and on the web you can download it as a JSON file.

The labels are defined in the code X published on GitHub. On a post, one reads "Post hidden from recommendations to non-followers." and another "Post is not shown on X." One account label is called DoNotAmplify: your posts are "hidden from recommendations to non-followers pending further review". It's the name the Twitter Files reported in December 2022, as "Do Not Amplify".

The report has limits. It covers the previous month only, and none of its label definitions mentions search, replies, trends or notifications. It has no button to appeal.

• A label on a post. Since April 2023, X's enforcement page says, posts that break its rules can get labels "informing both post authors and viewers that we limited the post's visibility." By November 2023 one read: "Visibility limited: this Post may violate X's rules against Hateful Conduct." X says: "Authors will be able to submit an appeal on the label".
• A limited account. If X limits you for aggressive following or engagement, you see "a message that your account's features have been limited for a specified time". In the meantime your account and posts "may be filtered out of certain places on X, including from search results and notifications." X says you can appeal "by contacting our support team".
• A label on your account. By late December 2023, users were reporting a notice: "We've added a temporary label to your account which may impact its reach." The ones we read gave spam or "platform manipulation" as the reason. In March 2024 TechCrunch found no way to appeal it; by December 2024 a user's screenshots showed a Request Review button. X's help pages don't describe the notice.
• In the EU, X's Digital Services Act page offers an appeal against decisions to "restrict visibility of information". On 5 July 2024, a court in Amsterdam ruled that X had restricted a user's account without the explanation the DSA requires.

Labelled posts lose most of their reach. In July 2023 Twitter said labelled posts "receive 81% less reach or impressions", that "More than one third of authors proactively choose to delete the post" once told, and that "only 4 percent of authors have appealed labels."

X search needs an account. When we tried on 11 October 2026, x.com sent a logged-out search to its login page, and X's help says advanced search is "available when logged in". So ask someone who doesn't follow you, or use a second account of yours that doesn't.

1. 1

   Search from:yourhandle in Latest

   Type from: followed by your handle, without the @, then open the Latest tab. X says Latest shows posts newest first, and that the only filter there removes "posts from spam accounts, protected accounts, deactivated accounts, and so on." If your recent posts are missing, that's a sign. One caveat: X's search FAQ says safe search applies "in both Top and Latest", so a post marked sensitive can be missing for that reason.

2. 2

   Type your name in the search box

   Check the suggestions that drop down, separately from the results. In 2018 Twitter found that some accounts "weren't auto-suggested in search even when people were searching for their specific name", though their tweets still showed in results. "Hundreds of thousands of accounts were impacted by this issue."

3. 3

   Open a thread you replied to

   Have your helper open a thread you replied to and look for your reply. X can tuck replies behind a button that, its help page says, "reveals replies that are identified as potentially lower quality and spam, based on factors such as the author's previous behavior and the content of the reply." The page calls it "Show probable spam".

4. 4

   Check your own account

   X's search help asks: "Have you confirmed your email address?" and "Is your name and bio filled in?" Look for a limit message when you log in, and for labels on your posts.

Some things look like a shadowban and aren't. On a busy post, X says, "our platform is unable to show all of these replies". A hashtag that's only a number, like #1, "is therefore not searchable". Notification filters can keep your replies and mentions out of the notifications of the people you reply to or mention. And X ranks some replies higher for a fee: the Premium page lists "reply prioritization", larger on each tier.

They test what anyone can see from outside: search and replies. None of them can see the labels in Under the Hood or how X ranks For you. On 11 October 2026 we gave four of them a handle that doesn't exist, @zqxnotreal98765:

A clean result from such a site proves little, and neither does a red one. The best-known early checker, shadowban.eu, online since at least July 2018, said on 2 December 2021 that it was "down for good": "The test results are not reliable, anymore. Twitter has changed their frontend to use new GraphQL endpoints."

X names these causes in its search rules, spam policy and automation rules, and Twitter named account signals in 2018:

• Following and unfollowing in bulk: "following and/or unfollowing a large number of unrelated accounts in a short time period, particularly by automated means", and "follow churn". More in how to mass unfollow on X.
• Hashtags: "posting with excessive, unrelated hashtags in a single post or across multiple posts", and abusing trending topics. X gives no number.
• An app posting the same content to many accounts: "Your account may be filtered from search if you are logged into a third-party application that is updating many accounts with similar or duplicate content."
• An account that looks new or fake. Twitter's 2018 post listed signals such as "whether you have a confirmed email address, how recently your account was created, whether you uploaded a profile image".

X says: "We do not block, limit, or remove content based on an individual's views or opinions." And if search hides you, "rest assured that your followers will still receive your updates." X gives no duration for any limit, apart from "a specified time" for a limited account and the "temporary" account label.

Rarer than it feels, in the two large studies, both with data from 2020 and 2021, before Musk bought Twitter:

• A study of a random sample of about 25,000 US accounts, published in the Journal of Communication in 2023, concluded: "In general, shadowbans are rare." Bot-like accounts were more likely to get one and verified accounts less likely. In a 2022 preprint of it, 6.2% of the 27,718 active accounts tested got at least one over a year; most of those were replies pushed behind a button (5.33%), and search bans were 0.91%.
• A 2021 study crawled more than 2.5 million accounts in April 2020, in the networks around chosen accounts. The share shadowbanned ranged from 0.50% around the French members of parliament to 2.34% around older accounts picked at random. The authors found the explanation that bans were bugs "statistically unlikely".

The first study found that replies from accounts posting offensive tweets or about politics, "from both the left and the right", were more likely to be pushed down. In a 2021 preprint, the same authors wrote that, controlling for incivility, bot-like behavior and network size, "there is little to suggest that right wing accounts are indeed more likely to be shadow banned."

Musk promised a way to see your account status in December 2022. What came was piecemeal: labels on posts in 2023, an account notice users found in 2023 and 2024, and Under the Hood in 2026, whose labels never mention search.

Every behaviour X names as a cause is something done from your account: following, unfollowing, posting, engaging, or an app posting duplicate content for you. UnfollowMonkey does none of it. We read your followers with our own access, and the only thing we ever write on X is a DM from @unfollowmonkey to you. Apps that do follow and unfollow for you are another matter: why X seems to unfollow people for you and is an unfollow tracker safe to connect?

We don't detect shadowbans either. What we track is your followers: who left, and whether X suspended them or they deleted their account.