The phone can replace the biometric machine — if the fence is honest. Here is the test.

Why schools are asking about geo-fenced teacher attendance

In January 2023, Jharkhand's school education department started tracking government-school teacher attendance with geofencing: the teacher's phone had to be inside the school's boundary for the day to count. Karnataka runs a similar Geo Attendance app for teachers registered in its SATS system. Private schools noticed. A principal in Gorakhpur who has spent three years nursing a fingerprint reader at the main gate — the PC that must stay on, the vendor's bridge that dies after every Windows update, the monsoon fortnight on paper — now asks a fair question: if the government can run teacher attendance on a phone and a circle, why am I still buying hardware?

The honest answer is that a phone can do it, and do it better, but only if the software behind the phone refuses to be lied to. That is the whole subject of this guide.

The thesis

A phone can replace the biometric machine for staff attendance in an Indian school — if, and only if, the fence is honest. Honest means four things: the punch time comes from the server, not the phone; a punch from a fake-GPS app, a stale fix or a weak signal is refused, not accepted with a note; every refusal is recorded where the office can see it; and the late mark is computed from the department's timings, not argued in the principal's office. A geo-fenced app that does less than this is a signature register with a map on it.

What geo-fenced attendance actually is

The school places a pin on a map — the gate, the staff room, the office — and sets a radius in metres. When a teacher taps Check in, the phone reports where it is and how sure it is (every phone reports an accuracy figure in metres alongside the coordinates). The server measures the distance from that point to the pin and accepts the punch only inside the circle. That is the fence. Everything else that matters — whose clock stamps the punch, what happens to a mock location, whether a refused tap leaves a trace — is a design decision the vendor made, and it is where honest and theatrical products diverge. A real implementation has all of the following, and you should be able to see each one on screen at the demo.

The India bar: where geo-fencing goes wrong in real schools

The failures are not exotic. A budget Android phone in a Tier-2 town reports 10 to 30 metres of accuracy outdoors and 50 to 100 metres inside a concrete staff room, sometimes worse on a cloudy morning. A school that sets a 50-metre circle at a strict accuracy limit will refuse honest teachers every day and quietly go back to the register within a month. A school that sets a 2-kilometre circle to avoid complaints has built a fence that includes the teacher's house. The right radius is a decision — 150 metres around the entry point suits most compact campuses — and the right product shows you, from day one, who was refused and from how far, so you can widen or tighten with evidence instead of complaints.

The second failure is trust in the wrong direction. Some products accept an out-of-circle punch and mark it "outside campus" for the admin to sort out later; within a fortnight nobody sorts anything out, and the flag becomes decoration. A fence that does not refuse is not a fence. The third is the quiet accept: a stale fix, an emulator's empty coordinates, a mock location — accepted because the app never asked. The fourth is buddy punching: a fence stops a teacher punching from home, but not a colleague carrying two phones through the gate. An optional selfie with each punch closes that for schools where it is real; face recognition is not required, a photo an admin can look at is enough.

How to test it at the demo

Do not evaluate a geo-fenced attendance app from a slide. Set it up on your own campus map during the demo and run these eight checks with your own phone. Each takes under two minutes, and together they separate an attendance record from a register with a map.

1. Set a 100-metre circle and tap from 300 metres away. The punch must be refused, the app must tell you the distance, and the admin screen must show the attempt with that distance. If it accepts and merely flags it, stop here.

2. Install a free fake-GPS app and set your location to the gate. The punch must be refused as a mock location, and the attempt must appear on the refusals list. A product that cannot see this flag has no answer to the cheapest trick there is.

3. Set your phone's clock back by an hour and check in at the gate. The recorded time must be the real time. If the punch shows the phone's time, the salary sheet can be edited from the phone's settings screen.

4. Check in inside a concrete room with weak signal. Watch what the app does with a fix that is only good to 200 metres. It should refuse as weak signal or wait for a better fix — not accept a guess.

5. Tap twice quickly, then try to check in a second time before checking out. Both must be refused with a clear reason, and the day must still have exactly one check-in.

6. Ask for the late math. Set a department in time of 8:00, a late-coming time of 8:15 and 10 minutes of grace, check in at 8:26, and confirm the teacher sees "Late by 26 min" and the register shows the same. Then excuse the late mark as an admin and check out in the evening: the excusal must survive.

7. Ask what happens to your fingerprint machine. A machine check-in at 7:50 should make the phone offer check-out, and the day should have one arrival, not two. If the phone and the machine keep separate registers, your accountant will reconcile them every month.

8. Ask who can see the selfies and the locations. The answer should be a named permission your school controls, storage in India, and no tracking between punches. Under the DPDP Act 2023 you are the data fiduciary for your staff's location data; make sure the vendor has thought about that before you have to.

The options you will run into

Two kinds of product answer this search. The first is the general-purpose GPS attendance app built for field sales teams and factories — names you will see include Truein, ubiAttendance, factoHR and factoTime, Timelabs, DigiSME and SalaryBox. They do geofencing and selfie punching well, are priced per employee per month, and know nothing about a school: no departments with in-times and grace periods, no link to the daily attendance board, no late tracker feeding a school salary summary, and a second login for staff who already use a school app. The second is the school ERP with a staff attendance module — Teachmint, Entab CampusCare, MyClassboard, Edunext, Vidyalaya and others. Most of them take biometric punches; fewer offer a geo-fenced check-in from the teacher's own phone, and among those that do, the honesty checks above vary widely. Ask each vendor the eight questions rather than reading the feature list.

The cost reality

A fingerprint reader is not the expensive part. The expensive part is everything around it: a PC at the gate that must stay switched on for punches to reach the software, the vendor's bridge that needs reinstalling after updates, an annual maintenance visit, a spare unit for the monsoon, and the fortnight on paper every time any of those fails. Multiply by three gates on a large campus. A phone-based system has none of this hardware and none of the reinstalls, and a new teacher who joins in July has the button on the same morning with no enrolment session.

Standalone GPS attendance apps charge per employee per month, on top of whatever the school already pays for its ERP, and add a second app to every teacher's phone. Inside a school ERP the feature usually costs nothing extra because it rides on the staff attendance module the school already has; check the pricing page of whichever product you evaluate, and ask specifically whether mobile check-in, selfies and the refusals list are in the plan you are buying or in a higher one.

Where Inkwelly fits

Inkwelly's Mobile Check-in is built inside its employee attendance module, so it is the same daily record, late tracker and salary summary the school already uses, not a second register. The school draws one circle and sets a radius between 50 metres and 5 kilometres; the server stamps every punch with its own clock, refuses mock locations, stale fixes and weak signal, allows at most 30 metres of slack for the phone's reported accuracy, and writes every refusal to an Activity screen with the distance, the accuracy, the device and the reason. A selfie can be required per school and is stored as a photo for attendance admins, not as face data. Mobile and fingerprint punches merge into one day, and late marking from department timings now runs the same way for every source. Location is read only at the tap, the data stays in India, and departments such as Transport can be excluded so drivers keep using off-campus requests. It is off by default; a school switches it on when it has drawn the circle.

“A geo-fence is not a feature. It is a promise that the software will say no — to the phone's clock, to the fake-GPS app, to the tap from the road — and write down that it did.”

Decide in two weeks

Pick one gate, one department and one fortnight. Draw the circle at 150 metres, leave the fingerprint machine running beside it, and ask ten teachers to use the phone. On day three, read the refusals list and widen or tighten the circle from what you see, not from what you hear. On day fourteen, put the late tracker next to the register the head clerk kept in parallel. If the two agree and the refusals list explains every gap, you have an attendance record that can end an argument. If they do not, you have learned it for the price of a fortnight, and the machine is still at the gate.

Run the eight tests on your own campus

A 20-minute walkthrough: we drop the pin on your school, set the radius, and you check in — and try to cheat — from your own phone.