AI & privacy

Student privacy and data handling

How student work and identity are protected, by architecture rather than policy alone.

InteGrade is built for institutions with real privacy obligations, and several protections are structural: the system is designed so certain data cannot leave your device, not merely promised not to.

Identity stays local

The automatic matching that links scanned submissions to your roster runs entirely in your browser. The handwritten names are read on your machine, compared against your roster on your machine, and the identifying page content involved in matching never leaves your device or reaches any outside service.

Names are removed before AI analysis

When an AI feature does process submission content on a server, such as course method analysis, student name areas are blacked out from the page images first, using the same name region you drew on the template. The analysis sees the math, not the identity.

Where data lives

  • Submission files and grades are stored in InteGrade's database and storage, scoped to your course. Access requires your sign-in, which runs through your Canvas identity.
  • Your in-progress grading is also held locally in your browser as part of the local-first saving model. Clearing browser data on a machine with unsynced changes can lose those changes, so let the sync indicator show synced before clearing anything.

FERPA

InteGrade operates under the school official exception to FERPA, processing student records solely to provide the grading service to your institution. Student data is never sold, never used for advertising, and never used to train AI models.

Questions

For security documentation, data processing details, or anything your IT office needs, contact support@integrade.app.