X's Under the Hood: the official report, and what a live check adds

Is X's Under the Hood a shadowban checker?

Not for today. Under the Hood shows the labels X applied to your own account and posts during the last calendar month. It is published about 10 days after that month ends, and only for accounts at least 1 year old with at least 10 posts that month. If Under the Hood looks clean and a checker finds a restriction, both can be right: Under the Hood covers a month that has already ended, the check covers today.

What Under the Hood shows

Under the Hood covers one calendar month at a time. Per X's published code, a month's report appears no earlier than 10 days after that month ends: from the 1st to the 9th of a month, the report still covers the month before last.

The report lists:

• Labels applied to specific posts, with the share of that month's posts carrying each one.
• The posts behind each post label, since October 30, 2026. In the default values of X's published code, the report lists every post when 3 or fewer posts carry the label, none when more than 5% of that month's posts carry it, and otherwise the 50 most recent. Posts withheld because of a legal requirement are always listed, up to the 50 most recent. X does not publish its live settings.
• Labels applied to the account as a whole, with the share of days in the month each one was active.
• Since September 19, 2026, whether your account or posts had their visibility limited because of what X calls "government-required content filtering".

X's code describes Under the Hood as a "transparency report". Since September 24, 2026, X also shows it on a "View report" page, and the JSON download is still available. The report opens at x.com/i/under_the_hood.

What each label does, in X's own words

X's published code gives every Under the Hood label a one-line effect. These are X's own sentences, read on November 5, 2026 in X's label file (commit 021afa4).

The effects that matter most, in X's wording:

• Likely spam, on a post: "Post hidden from recommendations to non-followers."
• Abuse and harassment, hateful conduct, violent content and civic integrity, on a post: "Post discoverability is restricted to the author's profile. All users can see a label explaining that the post has limited visibility."
• Insults, on a post: "Post hidden from recommendations to non-followers. All users can see a label explaining that the post has limited visibility."
• A post pending review: "Post is not shown on X while pending further review." A post pending deletion by its author: "Post is not shown on X while pending deletion by the author."
• Likely spam, on the whole account: "The account's posts are hidden from recommendations to non-followers."

Where the text says "non-followers", the posts stay visible to the account's followers. X's README says the extra rules apply only to a recommendation from an account the viewer does not follow, and that "the same post is allowed to a follower".

To contest a post label, X's report says, for the abuse, hateful conduct, violent content, civic integrity and insults labels: "If the post still has this label, you can appeal from the post."

For a post pending deletion by its author, the report points to the email X sent: "Instructions provided in email notification at the time of enforcement."

In the same published code, the automatic spam labels on an account last 7 days or 30 days, and one rule, for a spam campaign, sets no end date. The guide to X's temporary labels and limits covers the rules.

This is X's wording in its published code, not proof of what is switched on. X's live deployment switches are not published, so the code cannot show which labels apply to your account today.

Who can see Under the Hood?

Only the account's owner, logged in to X: X's code serves the report to the authenticated viewer, so no one can open another account's report.

The account must be at least 1 year old and have 10 or more posts in the prior month. Below either threshold, X shows an eligibility checklist instead of the report.

If X shows you that checklist, here is what to check instead.

What Under the Hood doesn't show

Under the Hood leaves out four things:

• Recent changes. The report comes out well after the month it covers. A restriction applied and lifted entirely since the last report was generated will not appear until the next one, which can be weeks away.
• The post behind an account-level label. The report gives the share of days that label was active, not the post that triggered it.
• An outside read. The report runs no search test, no thread visibility test and no ranking comparison. It reports labels X's own systems logged internally, not what search or an open conversation shows a visitor right now.
• Accounts below the thresholds. An account under 1 year old, or with fewer than 10 posts in the covered month, sees no label data at all in that report, regardless of what is actually happening to its reach.

What does a live checker add to Under the Hood?

Shadowban Radar, a free live checker, reads what X shows anyone today, on any public account. Compared with Under the Hood, it:

• works on any public account, including accounts under 1 year old or with fewer than 10 posts in the prior month;
• reads today's visibility, and every result shows when X's data was read;
• runs the 4 checks (search suggestion ban, search ban, ghost ban, reply deboosting) on X's public search results and conversations;
• gives a confidence level for each check, and marks a check Untestable when the evidence is missing;
• needs no login or password.

Shadowban Radar cannot read X's internal labels: Under the Hood is where X names them, for a past month. The method is on the methodology page.

Should I use X's Under the Hood or a live check?

X's Under the Hood shows the labels X applied to your own account and posts in a past month, if your account qualifies. Shadowban Radar complements it: it tests what other people see today on any public X account, for free and with no login.

• Under the Hood: X's official report, only for the logged-in owner of an account at least 1 year old with at least 10 posts that month. It comes out about 10 days after that month ends.
• Shadowban Radar: any public account, of any age, with no login. It runs the 4 checks (search suggestion ban, search ban, ghost ban, reply deboosting) on what X serves today, with a confidence level for each check.

If Under the Hood looks clean and a checker finds a restriction, both can be right: Under the Hood covers a month that has already ended, the check covers today. For today's answer on any account, run Shadowban Radar's free check; for the labels X applied in a past month, open Under the Hood if your account qualifies.

Sources

X's own posts and code, in date order:

• X's post introducing Under the Hood, @XOpenSource, August 13, 2026
• X's post expanding Under the Hood, @XOpenSource, September 19, 2026
• X's post on the easier report view, @XOpenSource, September 24, 2026
• X's post adding the list of affected posts, @XOpenSource, October 30, 2026
• X's published Under the Hood code, xai-org/x-algorithm, version of October 31, 2026
• X's file of label texts, xai-org/x-algorithm, commit 021afa4, read November 5, 2026
• X's README on recommendations and followers, xai-org/x-algorithm, commit 021afa4
• X's guide to posts with limited reach, help.x.com

To see today's visibility instead of last month's, test any public handle right now, free and in a few seconds.

The 4 restrictions, in detail

• Search suggestion ban
• Search ban
• Ghost ban
• Reply deboosting