The whole platform, module by module

Structured records with guardians attached, including the Nepal-specific and IEMIS fields institutions are asked to report.

Departments, designations, qualifications and documents, held in the same system as the people they teach.

Permissions come from a role, then extras and revocations per person, and every one of them is checked on the server.

Validate as a dry run, fix the flagged rows in a grid, commit, and revert the whole import if it went wrong.

Who changed what, when, and anything Gyanu drafted marked as AI-assisted rather than passed off as typed.

Set the structure once and carry it into the next year rather than rebuilding it.

Period, break and assembly slots, routines per class and per teacher, and one-off changes recorded as overrides.

Student and staff attendance, with teachers scoped to the classes they actually take.

One request flow for both. Approving a staff leave marks the attendance, so the two records cannot drift.

Homework set against a class and subject, and syllabus units marked off as they are covered.

Grading schemes, marks entry, re-exams, seat plans, invigilation duty, and results that stay private until published.

Preview, blockers, commit and undo, with each student's history carried forward.

Flexible cadence, invoices generated for a class, a batch or the whole institution, cancelled and regenerated with a reason.

Payments allocated against invoices, with a lifetime statement that reconciles.

A waiver applied once and applied the same way every time, and reversals that keep their trail.

What each student already owes comes across, so an institution switching mid-year does not start from zero.

Salary structures, monthly runs, SSF or provident fund, income tax bands, and payslips staff open themselves.

Issue five notebooks from the store room and they land on that student's bill without a second entry.

One record per title, one row per physical copy, and an accession number that is permanent and never reused.

Import the collection you already have, then print the labels on ordinary paper on an ordinary printer.

Scan, or search by name, or pick from the catalogue by hand. A copy can be on only one open loan at a time.

How many books, for how long, and how many renewals, because a young student, an adult learner and a teacher are three different loans.

Off by default. When a fine is raised, it posts into the same ledger as everything else, never a separate cash box.

A borrower places a hold on a title that is out, and the queue is honoured in order when a copy comes back.

Overdues show on the dashboard and each borrower's card, reminders go out on their own, and a shelf audit reconciles the catalogue against what is physically there.

Consumables alongside computers, projectors, furniture and sports equipment, with location, holder and condition.

Supplier, purchase order, approval and goods received, updating both the stock and the accounts.

Vehicles, drivers, routes, stops and student assignments, with a live map where compatible trackers are fitted.

Face, fingerprint or RFID terminals feeding staff attendance directly. Attendance works just as well without them.

Published to the roles and classes they are meant for, rather than to a WhatsApp group that half the parents left.

Standard for every institution. The in-app inbox is the record; the push is only the nudge that points at it.

A paid per-institution add-on for the institutions that need it, rather than a cost everyone carries by default.

One app that changes shape by role and enforces exactly the permissions the web dashboard does.

An inexpensive USB scanner at the desk, or the camera on a phone. Labels are generated and printed by NepEdu itself.

Questions, tests, homework, notes, lesson plans and slides, drafted from your own syllabus rather than from nowhere.

A drafted comment is never published on its own. The class teacher reviews and approves it first.

Works out which spreadsheet column is which when the deterministic rules cannot, so a messy file still imports.

AI is a per-institution switch, every call is logged with who ran it, and anything AI-assisted is marked as such.

Every record belongs to one institution and every query is filtered by it. If the institution cannot be established you get nothing, not everything.

Nightly, encrypted, held off the server, and restore-tested rather than assumed to work.

Records are soft-deleted and recoverable, because the most common data loss anywhere is somebody deleting the wrong row.

Bikram Sambat handled throughout, and the fields you are asked to report are fields you already hold.