Skip to main content
You can use your own API keys with HyperWhisper.
You do not need an API key to use HyperWhisper. By default, transcription runs on HyperWhisper Cloud. It works immediately after installation, and it needs no setup. Your first credit purchase (from $5) creates an account key with 5,000 credits (approximately $5 of value — see Providers). HyperWhisper Cloud is priced at face value with no markup. You pay the same per-minute rate as a direct holder of the provider’s API key.Use this page only if you want to send transcription or post-processing to your own provider account. For example, you have corporate credits, you have a free tier, or you want a model that Cloud does not offer.
HyperWhisper keeps your keys on your device. It uses each key only for requests to the API endpoints of that provider. For example, HyperWhisper sends an OpenAI API key only to OpenAI. It sends your keys to no other destination.

​
How to Get an API Key

These sections show how to get an API key from each provider.

​
OpenAI

​
Groq

​
Deepgram

​
Assembly AI

​
Soniox

Get your API key at console.soniox.com. [Video Tutorial Coming Soon]

​
Grok (SpaceXAI)

You can use Grok for transcription and for post-processing. Get your API key at console.x.ai. [Video Tutorial Coming Soon]

​
ElevenLabs

When you create your ElevenLabs API key, enable these permissions:
  • Speech to Text — necessary for Scribe transcription
  • Models — necessary for the model availability checks
Without these permissions, transcription fails, even when your API key is valid.
Do not set a credit limit on your API key. Keep the credit limit field empty. A credit limit of 0 blocks all requests, even when your account has credits.
[Video Tutorial Coming Soon]

​
Mistral

[Video Tutorial Coming Soon]

​
Anthropic

​
Cerebras

You can use Cerebras for post-processing only. Get your API key at cloud.cerebras.ai. [Video Tutorial Coming Soon]

​
Google Gemini

​
Gemini 3.5 Transcribe

Gemini 3.5 Transcribe has its own key slot, next to Google Gemini. Get the key at aistudio.google.com/apikey — the same page as Gemini.
Put the key in the Gemini 3.5 Transcribe slot. A key in the Google Gemini slot does not operate Gemini 3.5 Transcribe. The two slots are separate, because the two products use different endpoints of Google.

​
Meta Muse Voice Transcribe

Get a Meta Model API key from Meta for Developers. Add it to the Meta slot in Model Library → API Keys. The key stays in your operating system’s secure credential store. A direct Meta mode sends its normalized WAV file from your device to api.meta.ai. It does not send the request through HyperWhisper Cloud and does not use HyperWhisper credits. HyperWhisper does not make a hidden audio request to test the key. The app reports the key as configured and validates it on your first transcription. Gemini 3.5 Transcribe is a speech model. You use it for transcription and for streaming transcription. You cannot use it for post-processing. Use Google Gemini for post-processing. [Video Tutorial Coming Soon]