Platform / Data retentionSign inGet an API key

Data retention

Audio and transcripts are deleted on a schedule you control, per project. Deleting sooner is one API call.

The schedule

Each project has two settings, under Settings → Project in the console:

SettingMeaningDefaultRange
Keep audioHow long uploaded audio is kept after its last transcription finishes. Audio that is never transcribed is kept this long after upload.24 hours1–720 hours
Keep transcriptsHow long results are kept after a transcription finishes.7 days1–90 days

Audio is never deleted while a transcription that uses it is still running. When a transcript is deleted, the transcription record stays with content_deleted_at, and fetching the result returns 410 content_deleted.

Deleting now

curl
curl -X DELETE "$TWENTYSIX_BASE_URL/v1/transcriptions/$ID" -H "Authorization: Bearer $TWENTYSIX_API_KEY"

This deletes the transcript and, if no other transcription uses it, the audio — immediately.

What we keep

Usage records (duration, options, amount) and request logs (method, path, status, timing) are kept for billing and support. They never contain audio or transcript text.

Where it lives

Uploads, transcripts, usage records and logs are stored in the region of the project they belong to. Customer audio and transcripts are not used to train models without a separate written agreement.

Scriba is a product of 26AI Private Limited, Pune, India.© 2026 26AI Private Limited