No model training on uploads
Customer source documents are used to generate the requested training output, not to train foundation models.
Leap is built for reviewable SOP and policy training. This page gives buyers the short version of how uploads, generated outputs, subprocessors, deletion, and rollout approval are handled.
Customer source documents are used to generate the requested training output, not to train foundation models.
Customer source files, project records, and account data are scoped to the customer workflow. Access is limited to product operation and support.
Render-time narration, extracted images, and generated media may use unlisted public-read asset URLs so the video and SCORM pipeline can play them. Do not upload secrets that should never appear in training output.
The application pipeline runs on AWS infrastructure in us-east-1 with Supabase for application data.
Leap generates a draft package. Your SME, Quality, legal, or client reviewer stays responsible for approval before assignment.
Uploaded files and generated outputs are retained while the account is active. Project deletion is available inside Leap; full account deletion requests are handled within 30 days.
Leap uses Supabase, AWS, Vercel, Stripe, Resend, and AI infrastructure to operate the service. A current subprocessor and security overview is available before sensitive transfer.
Annual contracts, purchase orders, W-9 requests, vendor questionnaires, and LMS import review are handled through guided review.
Email sid@askleap.ai with subject [security]. For annual contracts, purchase orders, W-9 requests, or LMS import review, book a 15-minute call.