Public — Opt-in voice

Recipient-facing voice creation for opt-in (magic-link token)

Create a voice to collect the host's audio samples

Public, token-authenticated action that creates a new voice on the opt-in's podcast to collect the host's audio samples. The voice starts pending — upload samples and finalize it before it can be used or linked at opt-in completion. Requires prior code verification, otherwise returns 409. The created voice's URL is returned in the Location header.

Mark the voice ready once its samples are uploaded

Public, token-authenticated action that marks the voice READY once it has at least one uploaded sample; only a READY voice may be linked as the podcast's primary voice at opt-in completion. Returns 400 if no samples have been uploaded yet, and requires prior code verification, otherwise returns 409.

Synthesize a short preview of the cloned voice (returns MP3 audio)

Public, token-authenticated action that synthesizes a short sample of the cloned voice so the host can hear how it sounds. Requires prior code verification, otherwise returns 409. Keyless callers are rate limited per token and per client IP (429 on excess); a valid API key for the podcast's network bypasses the limit. Cloning and synthesis run against an upstream provider, which may return 502 on failure.

Upload audio samples for the voice (replaces the current set)

Public, token-authenticated action that uploads one or more audio files as the voice's samples, replacing any previously uploaded set. The voice stays pending until it is finalized. Requires prior code verification, otherwise returns 409.