More than 1,000 users for the AI transcription tool Whisper

In October 2025, the University of Groningen (UG) launched the transcription tool Whisper, which converts audio recordings into text. More than 1,000 UG staff and students are now using the free AI transcription tool. This rapid growth demonstrates that there is high demand for open-source AI tools hosted locally.

Transcribing an audio recording – such as an interview or a meeting – is a time-consuming task. Thanks to the rise of AI, there are now various transcription tools available for transcribing audio recordings. However, these tools often come with high licence fees, or there are concerns about data security and compliance with the GDPR.

Developed by the Digital Competence Centre (DCC)

The University of Groningen’s Digital Competence Centre (DCC) has developed a solution to this problem in the form of a locally hosted transcription tool based on OpenAI’s Whisper model. All researchers, staff and students at the University of Groningen can use this tool free of charge. The tool runs on the Hábrók High Performance Computing (HPC) cluster, which is managed by the University of Groningen’s Centre for Information Technology (CIT). This ensures that the data remains within the University of Groningen’s infrastructure.

A diverse range of users

Whisper can be used for a wide range of purposes, and its user base is consequently diverse: researchers, students and support staff all use Whisper. All faculties and departments at the University of Groningen are served, although it is notable that the majority of users are based in the Faculties of Behavioural and Social Sciences (red), Economics and Business (dark yellow), and the Faculty of Arts (blue). This differs from traditional HPC/AI applications, where users mainly come from the Faculty of Science and Engineering and the Faculty of Medical Sciences.

Secure alternative

Whisper runs entirely within the University of Groningen’s infrastructure on the Habrok HPC cluster at the CIT. Audio files are processed locally and are not sent to external AI providers or cloud services. This makes the service a secure alternative to many online transcription tools and suitable for processing sensitive data.