Proof of learning, not surveillance

The only thing AnTeraju records that matters is that a learner produced an answer, unaided, and it was signed as earned. The mastery ledger holds that result — a topic, a state, a timestamp — not a video of the student, not a keystroke log, not an engagement score. There is no camera-on proctoring and no attention metric, because the design grades what is made, not how long someone was watched making it.

A green star is a cryptographically signed entry in an append-only ledger, not a mutable number in a report. It says a topic was demonstrated; it does not say who the learner is to anyone without access. Teachers, schools and funders read the same record, and it is designed so a helper's author can see that mastery moved without being handed an individual's identity or history.

Dashboards for schools, ministries and funders describe cohorts, not children. Small cohorts are not broken out into identifiable slices, and no surface exports an individual learner profile. The procurement question — which helper raised mastery per ringgit — is answerable entirely from aggregate, signed state.

AnTeraju runs in Bahasa Melayu, English, Mandarin and Tamil, aligned to the national syllabus. The language a learner chooses is used to serve the right materials — it is never used to build an advertising or behavioural profile, because AnTeraju has no advertising layer to feed.

During the private beta run with QynQue, participating schools and classrooms are named only with their consent. Pilot results are reported without identifying any individual learner, teacher or institution that has not agreed to be named.

To exercise access, correction or deletion rights, or to ask how any field above is implemented, write to anteraju@antrex.app or use the Antrex contact desk at antrex.app/contact.