Skip to main content

Overview

Poll the status of an asynchronous text-to-speech (TTS) job. Use the runId and model values returned by POST /api/tts when the initial response is 202 with status: "pending". For low‑latency, synchronous TTS without polling, use Speech (POST /v1/audio/speech).

Query parameters

  • runId (string, required): Job identifier from the submit response
  • model (string, required): The model that was used for the job (e.g., Elevenlabs-Turbo-V2.5)
  • cost (number, optional): Cost from the ticket; helps with automatic refunds
  • paymentSource (string, optional): Currency/source from the ticket (e.g., USD)
  • isApiRequest (boolean, optional): Pass true when polling from API clients
Including cost, paymentSource, and isApiRequest from the original ticket helps the platform perform automatic refunds if the upstream provider rejects content after you were charged.

Usage

Response examples

Pending

Completed

Error (generic)

Error (content policy)

Notes

  • For Elevenlabs‑family async models (e.g., Elevenlabs-Turbo-V2.5, Elevenlabs-V3) you will always poll this endpoint until it returns completed.
  • When the job completes, the response includes an audioUrl you can download or play in the browser.
  • If available, include cost, paymentSource, and isApiRequest from the original ticket while polling to improve refund handling.