Everything you need to know about using Ultramic
Ultramic is a speech-to-text transcription assistant that works across all your desktop applications. It offers two modes: Cloud (OpenAI) for maximum accuracy and Local (Whisper) for privacy and offline use.
We believe in a pay-as-you-go model for cloud usage, so you only pay for what you use. No subscriptions, no hidden fees. Bring your own API key and you are ready.
Ultramic gives you the freedom to choose how your audio is processed:
To use Ultramic with OpenAI, you need to get your API key. Follow these steps to get your own key:
Note: Keep your API key secure and never share it with others.
We've designed Ultramic with a pay-as-you-go model instead of a subscription for several reasons:
OpenAI's Whisper API is very cost-effective, typically costing just pennies per hour of transcription.
If you prefer not to send data to the cloud or want to use Ultramic offline, you can use the Local Whisper provider. This runs the transcription model directly on your Mac.
To enable it, simply select "Local Whisper" in the application settings. The first time you use it, Ultramic will download the necessary model file (approx. 150MB - 500MB depending on the model quality selected).
Note: Local transcription is resource-intensive. It works best on Macs with Apple Silicon (M1/M2/M3 chips).
Ultramic uses keyboard shortcuts to make transcription easy and accessible. You can choose from two available shortcuts:
Option+spaceThe shortcuts work globally across your system, so you can use them in any application where you need to input text.
Ultramic simplifies the process of converting speech to text. Here is a typical workflow:
Option+space by default) to start recording.The entire process happens quickly, allowing you to write up to 3× faster than typing.
Ultramic supports 98 languages, including English, Spanish, French, German, Czech, and many more, thanks to the Whisper model (used in both Cloud and Local modes).
It depends on the provider you choose:
Ultramic is designed for short to medium-length dictations. For very long recordings, you might want to use a specialized transcription service.
Try speaking clearly and at a moderate pace. Ensure you're in a quiet environment with minimal background noise for best results.