Feature Request: Reuse uploaded file to branch new transcription jobs #7
Description
Background
Currently, each uploaded file can only be transcribed once. Users may want to try different settings (e.g. another model, with/without diarisation or translation) on the same file without uploading it again.
Requested functionality
• Introduce a new API endpoint (e.g. POST /jobs/{job_id}/branch)
• Allow the user to create a new transcription job using the same uploaded media
• The new job can specify new parameters:
• Different model (e.g. KB Whisper Large vs standard Whisper)
• Diarisation on/off
• Translation on/off
• Anonymisation on/off
• The new job should be treated as a separate job in the system, but reference the same file in storage
Expected behavior
• No duplicate file storage
• Full isolation between jobs (outputs, status, metadata)
• All job actions (view, edit, export) apply independently to each branch
Priority
High – improves flexibility, user control and avoids redundant uploads.