Guides
Safety, reports and acceptable use
RPT AI Video Studio has two safeguards: anyone signed in can report content, and a content check runs before paid work starts. The rules about what is allowed are in the acceptable use policy.
Acceptable use
What you may make and share is set by the acceptable use policy. Reports and the content check apply that policy.
Report content
A signed-in person reports content from the report dialog. The content can be a video, a project, an Explore item or a template.
| Reason | Use it for |
|---|---|
csam | Sexual content involving minors |
ncii | Intimate images shared without consent |
impersonation | Pretending to be someone else |
illegal_other | Other illegal content |
sensitive | Sensitive content |
copyright | Copyright |
spam | Spam |
other | Anything else |
- A note can be up to 1000 characters.
- You can send up to 20 reports per hour; more fails with
RATE_LIMITED. - You can report content in a workspace you belong to. Templates are shared by everyone. Anything else answers
NOT_FOUND. - A report needs a signed-in person in the web app. An API token or an MCP session gets
FORBIDDEN. - If an open report already exists for the same content, you get that report back instead of a second one.
- The owner of the content never sees who reported it.
What happens to a report
Our team reviews the report and either dismisses it or acts on it. Acting depends on what was reported:
- An Explore item is hidden. The project behind it is not touched.
- A template is unpublished.
- For a project or a video the decision is recorded; our team may also suspend the workspace.
Content check before paid work
Before a paid job starts, the text of the request is checked. It is on by default.
- A blocked term refuses the start with CONTENT_BLOCKED (HTTP 422). Its
detailsname thecategoryand thefield_pathof the text. The refusal comes before any hold or charge, so nothing starts and nothing is charged. - Other terms only warn. The work goes ahead and the reply carries
content_warnings, a list ofcategoryandfield_path. A warning never blocks delivery.
For an agent, CONTENT_BLOCKED has the retry policy after_fix: change the text named by field_path, then quote again. Do not send the same text twice.