Voice AI of KISSKI and the GWDG

In the Academic Cloud you can find an AI-Tool usable for transcibation and translation of audio files. You can access via the AcademicCloud or alternatively directly using this Link. You can easily log in via SSO and start transcribing your very own audiofiles.

The Voice AI can also be used when working in multilangual scenarios. It offers great support by providing translations. As an interesting asset you can even use the VoiceAi to directly generate subtitles for presentations or meeting in BigBlueButton. These are provided live and can also be provided in different languages.

User Interface

By pressing "DATEI AUSWÄHLEN" you can easily upload your own files into the AI interface (see picture below). Your own files are not saved on any databases during this process.

There are many different languages available and different text formats can be assigned as wished output format. You can choose between a normal textformat and formats especially provided for the creation of subtitles. The format SRT should be used when creating subtitles for videos, while the format VTT can be used when specifically creating subtitles for webvideos. VTT supports metadata and is often used in HTML5-Videos.

Combining Voice AI and BBB

A special feature is the combination of the Voice AI and the conference tool BigBlueButton (BBB). Here you can access the transcribation settings directly when entering a conference. Die Textausgabe findet sich dann im gemeinsamen Chat. Informationen zur Nutzung von BigBlueButton finde Sie unter diesem Link

Data security

For the usage of the Voice AI no private data is stored - neither the uploaded data, nor live transcripted audio inputs. There are no talks or data used to train this or other AI models.

For questions or feedback please feel free to contact us or directly contact support@gwdg.de.