aTrain is a desktop application that automatically transcribes speech recordings, such as research interviews. It offers a user-friendly interface to the faster-whisper implementation of the Whisper model, supports many languages and has an optional speaker detection mode based on pyannote.audio. The transcripts can be imported into the qualitative-analysis tools ATLAS.ti, MAXQDA and NVivo. aTrain runs on the CPU or on a CUDA-capable NVIDIA GPU and is distributed through the Microsoft Store for Windows and through Flathub for Linux.
aTrain was developed by researchers at the Center for Data Science in Business and Society (BANDAS) at the University of Graz in Austria. The project itself states that it is not an official app of the university. The source code is published under the AGPL-3.0-or-later licence, and the project is listed in openCode, the German software directory for the public sector. According to the project, all processing happens offline on your own device: recordings and transcriptions are not sent to the internet, so the data stay with the user.