DDLS Transcription

Drop in an audio/video file or record from your mic. Runs on the course Whisper (whisperX, large-v3) GPU service. Free for course use.

Click to choose or drag & drop audio/video files here

mp3 · wav · m4a · webm · mp4 · ogg — you can queue several

Prefer to transcribe with your Pi agent? (final-project students)

If you are on a final project, your project has a built-in transcription endpoint that runs on our server and saves the transcript straight onto your project. It is a good option for long recordings. Ask your Pi agent to read your project's skill document and transcribe the file, or run it yourself with your DDLS key (the same one you use for the model gateway):

curl -s -X POST https://ddls-portal-6228434e.svc.hypha.aicell.io/api/project/<your-project-id>/transcribe \
  -H "Authorization: Bearer $DDLS_API_KEY" \
  -F audio=@your-recording.m4a

Find your project id in the portal (it is the number in your project page URL). The reply contains the transcript text, and it is also saved to your project automatically.

Audio is sent to the course GPU service for transcription and not stored by this page.
Shared GPU — please keep files reasonable and don't run many at once. Questions: ddls-course@scilifelab.se