Norwegian version of this page

Autotekst – speech to text with Whisper from OpenAI

What is Autotekst?

Autotekst is a UiO developed tool for automatic transcription of speech to text powered by Whisper from OpenAi.

The service runs entirely on UiO servers, thus no data ever abandons the UiO infrastructure. This way the service can be used for certain types of red data, if your area of work requires you to do so.

Illustration giving an overview of data-flow in Autotekst

Overview of data-flow in Autotekst

Autotekst can also be used to transcribe audio files, like recordings done with the service Nettskjema-diktafon, Zoom or Panopto.

See also video demonstration of the new version of Autotekst.

Get started

Log in to Autotekst?

What can be sent and stored in Autotekst?

Up to red data

Guidelines for using Autotekst

  • Free usage for UiOs employees and students.
  • The service should only be used for data?owned by you or the institution you work for.
  • Don't hesitate to contact us if you plan to auto-generate text to?a large amount of sound files.
  • The service can be credited with the following text: “This work was carried out with the help of Autotekst. Autotekst transcribes speech to text using Whisper from OpenAI. The service is owned by the University of Oslo (UiO), and is developed and operated by the IT Department at UiO (it-hjelp@usit.uio.no)."

Help and instructions

S?k i veiledningene