FasterWhisper Audio Transcription
What It Does
This app turns an uploaded audio file into written text. It supports .mp3, .flac, and .wav files, and lets you choose between several Whisper models, from small and fast models to larger multilingual ones.
After transcription, the app shows the text in timed segments. You can listen back to each segment, correct the text directly in the browser, and download the final transcript as a text file.
When To Use It
Use this app when you already have an audio recording and want a first transcript that you can check, correct, and reuse in your research.
Research Examples
- Transcribe recorded oral history interviews so you can search, code, or analyse the interview material.
- Create draft transcripts of lectures, focus groups, podcasts, or spoken-language data before manual correction.
Good To Know
The first time you use a model, it may take longer to load because the model has to be downloaded and cached on the HPC system. After that, the same model should usually load faster.
You can find this app in the B-TXT toolsuite under Interactive Apps on the VUB-HPC portal.