X is testing a feature intended to show users whether its algorithm has limited the visibility of their posts, addressing long-standing claims of “shadowbanning”. The tool is part of a broader effort to increase transparency around X’s recommendation systems, although its usefulness may be limited by technical complexity and incomplete disclosure.

Available initially to a small group of testers through an “under the hood” settings page, the feature provides aggregate data from the previous month rather than per-post results. Users download a JSON file showing visibility-impacting labels, such as NSFW, spam, violence, impersonation, hateful conduct and civic-integrity flags; some labels can hide posts from non-followers or restrict them to the author’s profile. X has released more ranking-algorithm code but continues to withhold material it says could enable manipulation, including advertising-related systems.