PostEaseThe short answer
A labeler is a moderation service on Bluesky. It attaches labels, short tags such as spam or graphic-media, to posts and accounts, and people who subscribe to it choose what each label does in their own app: hide the content, show a warning, or ignore it. Bluesky introduced them in 2024 as “independent moderation services,” which you subscribe to in the app much as you follow an account.
Every Bluesky app user is subscribed to Bluesky’s own moderation service by default. Labelers add layers on top: Bluesky calls this a “stackable ecosystem of services,” where you install filters from independent services “like building blocks” over the app’s foundation.
How a label works
Bluesky’s developer documentation describes the idea behind it: speech and reach are separate layers, and moderation is implemented with labels. A label is metadata about a post or account, published by a moderation service, and it can be time-limited. Each labeler defines its own label values and how each one behaves:
| Setting | Options | What it controls |
|---|---|---|
| blurs | content, media or none | Whether the whole post, only its images, or nothing is hidden |
| severity | alert, inform or none | Whether a “danger” warning, a neutral information label, or no label is shown |
| defaultSetting | hide, warn or ignore | How the label is set when someone first subscribes |
| adultOnly | true or false | Whether the label can be configured with adult content turned off |
What you control as a reader
For each label from each labeler you subscribe to, you pick hide, warn or ignore. The protocol guide says hide and warn are similar, except that hide also removes the labeled content from feeds and listings. A few labels are defined by the protocol itself and cannot be overridden by any labeler:
| Label | Effect |
|---|---|
| !hide | A warning that cannot be clicked through, and the content is filtered from listings |
| !warn | A warning that can be clicked through |
| !no-unauthenticated | Content is not shown to logged-out users in apps that respect the label |
| porn, sexual, graphic-media | A warning on images, clickable only by people who are 18 or older with adult content turned on |
| nudity | A warning on images, not limited to 18 or older, ignored by default |
| bot | Marks an account as automated |
Who runs labelers
Anyone can. Bluesky open-sourced Ozone, its labeling tool, so a person or team can review reports in a queue and label posts or accounts, and labels can also be applied automatically through the API. A labeler runs from its own account and can accept reports through the in-app reporting flow, so a community’s moderators do not have to be tagged every time. Bluesky notes labelers work across the AT Protocol network, not only in the Bluesky app, and that nothing stops one from having paid subscribers.
What it means for creators
A labeler’s label follows your post wherever its subscribers see it. If your posts are being hidden or blurred for some readers, a labeler they subscribe to may have labeled them. Accurate labels on your own sensitive content help too: the protocol lets a poster apply a few global labels themselves, such as porn, sexual, gore and nudity, so readers get the warning they chose instead of a surprise.
Questions
How do I subscribe to a Bluesky labeler?
Open the labeler’s profile in the Bluesky app and subscribe, the way you would follow an account. Then choose hide, warn or ignore for each of its labels.
Can I turn off Bluesky’s own moderation?
Not in the Bluesky app. Bluesky’s moderation service is the default for every app user; to opt out of it entirely, Bluesky says you would use or build a different client app on the AT Protocol.
Can anyone create a Bluesky labeler?
Yes. Bluesky open-sourced Ozone for reviewing reports and labeling content, and labels can also be applied automatically through the API. Some technical knowledge is needed.
What is the difference between a labeler and a moderation list?
A moderation list mutes or blocks the accounts on it. A labeler can label individual posts as well as accounts, accept reports, and let each subscriber choose what every label does.
Sources
- Bluesky, Bluesky’s Stackable Approach to Moderation, March 2024.
- Bluesky, Moderation (developer documentation), checked October 2026.
- Bluesky, AT Protocol: Labels, checked October 2026.