Memra is the reasoning layer your MIS never had — the brain to your school's management information system. Staff ask a plain question, Memra answers from real, retrieved records, cites the source, and never shows more than that user could already see natively.
No live data connection yet. Read-only, always. Built for iSAMS first — designed to extend to other school MIS platforms.
No live query hits the source system directly. Data is synced on a schedule, then every answer is retrieved and summarised strictly from what's already stored — nothing generated from nowhere.
Read-only batch pull from the MIS API, on a schedule — never a live hit per chat message.
Short retention, access-controlled, encrypted at rest. Only what's needed for the domains in scope.
Every lookup filtered through the asking user's own permissions — never broader than their native access.
Summarised strictly from retrieved records. Source referenced. Logged in the audit trail.
This isn't "AI on top of an MIS." It's designed around how school data has to be handled, not around what's technically possible.
Memra never writes back into a school's MIS. There is no write path.
Any real pupil data connection requires that school's own sign-off, a DPIA, and admin-issued API credentials — never a workaround.
Only the data domains actually needed — attendance, timetable, assessment. Nothing broader, no blanket scopes.
Every part of this is built and proven against synthetic or sandbox data before any real pilot is considered.
No live pupil data. No school connection. Full transparency, on purpose.