If two permitted viewers see different Bluesky content, start with the simpler explanation: each account can layer different moderation services and filters, so a label, warning or hidden item can look different from one screen to another. Record what each view actually showed, note which viewer-side setting could account for it, and keep author deletion, edits and unavailable material as separate possibilities until something concrete supports one of them.
Find public social profile clues with our TraceSeek app.
Start by describing the two views, not the person
The task is a comparison, not a judgment. Write down what each permitted viewer saw in their own session: whether content appeared at all, whether it carried a warning or label, and how the surrounding thread looked. Keep the two observations side by side instead of merging them into one story. Bluesky's baseline moderation can be combined with optional independent services whose labels and filters are layered into the viewer experience, which means a difference between two screens is often an expected product behavior rather than a contradiction. This is the first place to look before anything else is concluded. [Bluesky: stackable moderation services]
Separate viewer-side moderation from author-side removal
A warning or missing item seen by one viewer does not by itself prove the author deleted anything. Because different selected services and settings can change whether labeled content is displayed, a viewer-dependent warning or hidden item is not equivalent to author deletion. Treat these as distinct categories with separate evidence needs: viewer-selected filtering, author-controlled reply tools, and genuine deletion of a post. If the only observation you have is that one view differs, you have support for the first category and not the others. Keep that distinction in your notes rather than smoothing it away in a summary. [Bluesky: stackable moderation services]
A small, honest record beats a confident guess
The table below is a reader worksheet, not a description of any app control. Fill in only what you or another permitted viewer actually observed, and leave blanks where nothing was seen. A blank is a finding too: it marks the point where an inference would stop being supported.
| observation | limit | next step |
|---|---|---|
| One viewer sees a warning label on a reply; the other sees the reply plainly | Does not show which service or setting produced the label | Note both views as seen; do not conclude the reply was deleted |
| A reply is absent in one thread view but present in another | Does not establish the cause | Check whether the difference is consistent with viewer-selected filtering before considering other explanations |
| No content is visible to either permitted viewer | Could reflect many causes, including account privacy or access limits | Record the absence and stop; do not attempt access outside permitted settings |
Do not use a moderation difference as a fact about a person
A label, warning, hidden item or absent reply is a moderation outcome, not evidence about someone's honesty, intent or identity. It should never be read as proof that a person is lying, cheating or unsafe, and it should not be used to draw conclusions about finances, relationships or private life from visible material. The same applies to the underlying post: a public profile or image may support broad public-source findings, but it does not verify identity or current activity. Describe what the moderation layer did and leave character judgments out of the record. [Bluesky: stackable moderation services]
When the material is absent, private or ambiguous
If the source you were comparing is missing, restricted or unclear, say so precisely and stop there. Do not try to reconstruct it through private access, identification of another person, or workarounds that evade safety and access settings. An unresolved note is a legitimate outcome: for example, 'reply visible to viewer A, warning shown to viewer B, no further detail available.' Mark the question open rather than filling the gap with a theory. The reader's next decision should depend on what is actually known, not on what seems most likely. [Bluesky: stackable moderation services]
A calm reading order for the next comparison
Before repeating a comparison, note each viewer's own account state as they describe it, because different selected services and settings can change what is displayed. Then describe what each viewer saw without interpreting it. Then state what would need to be true for an author-side explanation, and whether any observation supports it. Only after that should you consider whether to respond at all. If a difference remains unexplained, the practical choice is usually to withhold a conclusion and, where relevant, rely on the ordinary platform reporting or support options shown in your own app version.
- Record both views in the same session where possible
- Label each note as observation or inference
- Leave the cause open when no observation supports a single answer
A couple more questions
Does a warning one viewer sees mean the author hid or deleted the post?
No. Different selected moderation services and settings can change whether labeled content is displayed, so a viewer-dependent warning or hidden item is not by itself proof of author deletion. Treat the author-side explanation as a separate possibility that needs its own support.
What should I do when the material is no longer available at all?
Record the specific absence and stop there. Do not seek private access or try to identify anyone to fill the gap; an unresolved note is the honest result, and it keeps your later decisions grounded in what you actually saw.
Can a moderation label tell me something about who the person is?
It cannot. A label, warning or hidden item is a viewer-side display outcome, not evidence of intent, honesty or identity. Keep the observation about the content and the moderation layer, and leave judgments about the person out of your record.
Sources & editorial notes
Official references checked on October 24, 2026. Checklists, comparison tables and example wording are DateClues editorial tools, not official tests or validated risk scores.
General information. No guide or tool can certify a person’s identity, intentions or safety. Read our editorial standards.