To build a user interface focused on enabling efficient human correction of AI-transcriptions, powered by the open source Whisper model. Currently, the vast majority of interviews done by Swedish researchers include personally identifying information which forces researchers to manually transcribe their interviews. This can take up to 8 hours per hour of recordings. We will host a Whisper Model and connect it to a modern user interface so researchers can upload, transcribe, correct, and in the future even thematically code their interviews with the support of AI models (initially Whisper for transcription. Possibly language models for thematic coding in the future). We aim to cut the time to accurately transcribe an hour of interview down to 30 minutes. We are looking to deploy an initial proof of concept of our tool on this cloud (which allows us to show the tool internally at UU to gather more momentum behind it, in order to gather the resources to be able to deploy it in a safer environment and enable usage with personally identifying information)