#how do credits work for the speech to text realtime api?

1 messages · Page 1 of 1 (latest)

olive caveBOT
#
AI-generated Answer for andreistirb_2021

Short answer

  • Unit of billing: Speech-to-Text (including the Realtime WebSocket API) consumes audio credits based on the number of minutes of audio you process. ElevenLabs’ credit system treats “audio” workloads as per‑minute usage. (help.elevenlabs.io)
  • Realtime endpoint: You use Scribe v2 Realtime over a WebSocket (wss://api.elevenlabs.io/v1/speech-to-text/realtime). Credits are deducted for audio you stream for transcription. (elevenlabs.io)
  • Add‑ons that change credit cost: Enabling Entity Detection on Speech‑to‑Text increases cost by 30%. Keyterm prompting and other advanced STT features may add cost (see pricing). (help.elevenlabs.io)
  • Plan credits and overage: Each plan includes a monthly credit allotment (with up to two months of rollover). If you’re on Creator or above, you can enable usage‑based billing to buy extra credits at a fixed $/1,000‑credits rate (Creator $0.30, Pro $0.24, Scale $0.18, Business $0.12). (help.elevenlabs.io)
  • Monitoring: Track your consumed credits and minutes in Developers → Usage within your dashboard. (help.elevenlabs.io)

Notes

  • Exact per‑minute credit rate for STT is not listed publicly in the docs; it depends on plan and any enabled STT features. Refer to the pricing page for your plan for the current rates. (elevenlabs.io)

Docs cited

If you need plan‑specific credit-to‑minute figures

blazing raptor
#

actually I would very much like to know the exact per-minute credit rate, as in the pricing page you state the included number of hours, and in my account I have only credits. What is the link between them? Let's take into account only the speech to text realtime api