What can limit reach
Some posts are covered or removed after scoring.
A post can score well and still be limited for a specific viewer. Explore how public labels, settings, account state, and recommendation context can allow it, cover it with a warning, or drop it.
Try a visibility scenario
Change who is viewing and what labels are present.
Base home …
→Recommendations …
Three action families
A visibility verdict is not another score.
The outcome answers what may be shown and how. It does not add or subtract RankingScorer points.
No deciding rule removes or covers the candidate for this viewer and surface.
The candidate can remain in the feed behind a warning, cover, or other treatment.
The candidate is removed from this context, with possible effects on thread, quote, or repost presentation.
Labeling path
Many systems contribute facts. Rules decide the treatment.
Models and enforcement services generate signals that are aggregated into stored labels. Visibility rules then interpret those labels for a product and viewer context.
Open visibility packageCompiled content understanding node · source only.
Open media-model-proxy/README.rstCompiled content understanding node · source only.
Open agatha/scalding/AgathaFeaturesPredictJob.scalaCompiled labeling rules node · source only.
Open safety-label-user-agg/postToUserLabelRules.stratoThe official release has intentional gaps.
Grox prompts and some Botmaker rules are not published. This site names that boundary instead of filling it with guessed logic.
View source on GitHubInspect your own account and post labels.
Under the Hood exposes aggregate statistics about visibility-impacting labels on an account and its posts.
View official description Open official tool