Submit audio moderation
Submit an audio clip for moderation. The clip is transcribed and the resulting text runs through the same keyword blocklist and LLM prohibition check as prompt moderation. Pass a publicly reachable HTTP or HTTPS audio URL — the worker downloads it. URLs that use another scheme, or that resolve to a private or internal network address, are rejected with 400 INVALID_REQUEST. The clip must be no larger than 100 MB and no longer than 30 minutes.
This moderates transcribed speech, not non-speech audio content — music, sound effects and other non-verbal audio are not classified.
The transcript is never stored and never returned. A rejection reason is model-generated text, scrubbed of verbatim quotation on a best-effort basis, and may paraphrase what was said.
Returns a job ID for polling or webhook delivery.
moderate:audio scope. Requests without this scope receive a 403 FORBIDDEN
response. See authentication.reason may contain.