Under the Hood, explained.
Every month X checks your posts and your account against its label list and tells you what it found, in a report called Under the Hood. A label can hide your posts from everyone who doesn't follow you, and nothing in your analytics says so. Read yours here, see what each label does, and share the result.
X's monthly report lists every label its systems put on your posts or your account. Drop your report in station 01 to read it label by label and make a card you can post.
Your report
Download last month's report from x.com/i/under_the_hood and drop it here. This page reads it in your browser. Nothing is uploaded.
- Sign in to X and open x.com/i/under_the_hood. It's also under Settings and privacy, then Your account.
- Pick the month that just ended and choose Download.
- Drop the small text file in the box above.
Your result and your share card show up here.
For each label: what X says it means, what it did to your reach, the posts it was on, and how to fix it or appeal.
Making your card...
Post on X opens X's post box with your words and your card's link. X shows the card under the post. Download the card if you also want it on Facebook or LinkedIn.
Who sees a labeled post
The same post, sent to two people: one who follows you and one who doesn't. Pick a label. These are the effects X lists for each one.
How labels work
Ranking decides the order posts appear in. Visibility decides whether a post can be shown to someone at all. A label works on the second one.
The post goes to X's servers like any other.
Classifiers and user reports look for spam, risky links, adult content and rule breaks.
To one post, or to the whole account for 7 or 30 days.
Each time someone's feed is built, the rules read the label and decide who can see the post.
Early the next month, Under the Hood lists every label that was on you.
Why it matters: a label doesn't lower your post's score. Your followers keep seeing you, so likes from them look normal. The only sign is that impressions from people who don't follow you fall off. The Under the Hood report is the one place X tells you why.
Early the next month. August's report was built September 9, and September's on October 7.
A label that sat on a post for three days in August shows in August's report, even if it's gone now.
X calls the report best effort. It reports labels, not reach, so it can't explain a drop on a post with no label.
X's code builds a report only for an account more than 365 days old with at least ten eligible posts.
The report, line by line
It's a small text file. Select a line to see what it means. Since October 30, 2026, the web version also lists the labeled posts when a label is on three or fewer posts, or 5% or fewer of the month's posts, up to the 50 most recent.
- Go to
x.com/i/under_the_hood. It is also under Settings and privacy, then Your account, then Under the Hood, and it's linked from Creator Studio. - Pick the month that just ended and choose Download.
- Open the small text file and read
postLabelsandaccountLabels.
Each post label found, with how many posts carry it and their share of the month. Since October 30, 2026, X also lists the labeled posts themselves. Empty is what you want.
Every label X can put on you
The full list from the file X uses to build the report, in X's own words.
| Label | What it means | What it does |
|---|---|---|
SPAM_HIGH_RECALL | Automated systems think the post may be spam | Hidden from recommendations to non-followers |
SPAM | Automated systems or a user report found a likely Authenticity violation | The post is not shown on X |
MALICIOUS_URL | The post likely links to a malicious site | Hidden from recommendations to non-followers |
DO_NOT_AMPLIFY | Older version of the malicious-link label, being retired | Hidden from recommendations to non-followers |
PDNA | Likely terms-of-service violation, pending review | Not shown on X while pending review |
BOUNCE | Terms-of-service violation confirmed | Not shown on X until the author deletes it |
FOR_EMERGENCY_USE_ONLY | Incident-response label with a custom notice | Behind a notice; not in the Home timeline |
FOSNR_ABUSE | Abuse and Harassment violation, from a user report | Visible only on the author's profile, with a public "limited visibility" label |
FOSNR_HATEFUL_CONDUCT | Hateful Conduct violation | Visible only on the author's profile, with a public label |
FOSNR_VIOLENT_SPEECH | Violent Content violation, from a user report | Visible only on the author's profile, with a public label |
FOSNR_CIVIC_INTEGRITY | Civic Integrity violation, from a user report | Visible only on the author's profile, with a public label |
FOSNR_ABUSE_INSULTS | Abuse and Harassment violation (insults) | Hidden from recommendations to non-followers, with a public label |
BRAZIL_ELECTION_LEGAL | May contain AI-generated election media that breaks Brazilian law during the pre-election silence period | In Brazil, hidden from recommendations |
NSFW_HIGH_RECALL | May contain adult content | Hidden from non-followers, underage, no-stated-age and logged-out viewers |
NSFW_HIGH_PRECISION | Likely adult content, by automation or report | Behind a content warning; hidden from non-followers and the sensitive groups |
NSFW_TEXT | Likely sexually explicit language | Hidden from non-followers and the sensitive groups |
NSFW_CARD_IMAGE | Link preview image likely adult | Behind a warning; hidden from non-followers and the sensitive groups |
GORE_AND_VIOLENCE_HIGH_PRECISION | Likely violent content | Behind a warning; hidden from non-followers and the sensitive groups |
NSFW_ADMIN | Written by an account X marked as primarily adult or violent, after user reports | Behind a warning; hidden from non-followers and the sensitive groups |
SpamHighRecall | Automated systems think the account is likely to post spam | Every post hidden from recommendations to non-followers |
AbusiveHighRecall | Automated systems think the account may violate the Authenticity policy | Every post hidden from non-followers; a security challenge is required |
ImpersonationHighPrecision | Likely impersonation | Every post hidden from non-followers |
Compromised | The account appears taken over | Every post hidden from non-followers; password reset required |
ReadOnly | Locked read-only until a violating post is removed | Can only read; every post hidden from non-followers |
DoNotAmplify | Likely terms violation, pending review | Every post hidden from non-followers pending review |
NsfwHighRecall / NsfwHighPrecision / NsfwNearPerfect | May, likely, or almost certainly posts adult content | Every post hidden from non-followers |
NsfwAvatarImage / NsfwBannerImage | Adult profile photo or header | Every post hidden from non-followers |
NsfwAdmin | Marked by X after reports as primarily adult or violent | Every post behind a warning; hidden from non-followers and the sensitive groups |
No label matches.
Most people on X have never opened this report. Share the page and they can read theirs in a minute.