Skip to main content
Successful text-to-speech requests are recorded in history with input text, voice, status, timestamp, and audio. Async text-to-speech jobs use the same id for generation_job_id and history_item_id, so they appear in history while they move from active to ready or failed.

Listing recent requests

Downloading audio

Pull the original audio for a recorded request and save it locally. The logs page also offers inline playback and a direct download link. For async jobs, use the generation jobs audio endpoint. Active jobs return GENERATION_NOT_READY with Retry-After.

Deleting items

Purge test runs before sharing logs, or clean up old generations. The SDK exposes history.delete; the HTTP equivalent is DELETE /v1/history/{history_item_id}.

List history items

Page through recorded requests and inspect filters, fields, and SDK examples.

Download history audio

Pull the stored audio for any history item; async jobs use the generation jobs audio endpoint.

CLI history and jobs

List, play, download, and delete history items and async jobs with breeze history and breeze jobs.

Text to speech

Generate the sync, async, and streaming requests that appear in history.

Developer logs

Review request status, latency, playback, downloads, and Copy as cURL.

Errors

Handle not-ready jobs, missing history items, auth failures, and transient generation errors.