Hello Square team,
We are planning a seller-authorised appointment assistant using ordinary
REST APIs and OAuth for customer lookup and caller-confirmed booking,
rescheduling and cancellation. Our proposed design prohibits using Square
Content to train, fine-tune or improve models across customers, and requires
contractual restrictions on our AI suppliers before any such processing.
Could the appropriate team clarify these points under the current Developer Terms?
- Is passing the minimum necessary Square Content to an AI model solely for
inference to fulfil that seller’s request permitted under §§2.4 and 7? - After a seller revokes access, we would stop API access immediately. Under
§3, how should Square-derived fields in historical call records and booking
audit records be treated? Is any limited post-revocation export or retention
permitted, including records specifically required by law? - What deletion timing applies, including inaccessible encrypted backups
that expire on a bounded schedule and have deletions reapplied before restore?
If these questions need a private contractual review, please identify the
appropriate route. We are not requesting permission to train models or
claiming that ordinary API access overrides the terms.
Thank you.