Built for the way programming is really taught

Why this exists

In programming courses, “did this student do their own work?” is a harder question than it looks. Every submission starts from the same assignment, often the same starter code, and there are only so many reasonable ways to write a sorting routine. A rename here, some reformatting there, and copied work can look superficially original — while genuinely independent solutions can look suspiciously alike. Instructors are left squinting at diffs, and integrity offices are left with weak evidence.

Honor Checker was built to change that. It compares the structure of code rather than its surface, using a published, reproducible fingerprinting algorithm — so cosmetic edits don’t hide copied logic, shared scaffolding doesn’t become a false accusation, and matches can be shown side by side as evidence an instructor can actually stand behind.

Built by an educator, refined in real courses

Honor Checker is developed by someone who teaches, and it’s shaped by the realities of the classroom: batches of submissions, sections that share students, assignments recycled across terms, and the need for evidence that holds up in an integrity review. It’s currently in beta at Brigham Young University, where it’s being refined in real courses — a reflection of the product’s stage, not an institutional endorsement.

Where we’re headed

The beta is closed, and general availability is coming. Our commitment as the product grows is the same principle that guides its AI roadmap: deterministic evidence first, assistance second, and a human always deciding. Honor Checker helps instructors see and understand the evidence — it never renders a verdict about a student.