Labour attendance app
Site hazri that can prove itself on the 30th
Mark labour attendance on site with GPS and a selfie, with or without signal, and let the same record produce the wages, the overtime, the advances and the PF and ESI reports. Built for contractors and site supervisors in India.
What actually goes wrong with a labour register
- The supervisor marks the hazri in the evening, from memory, for a morning he was not there for
- On payday a worker says he did 26 days and the book says 24, and nobody can now prove either number
- Half-days and overtime are settled by argument, because the book records a fraction and not a time
- A worker does Monday and Tuesday at one site and Thursday at another, and two registers have to be reconciled by hand
- The mate is paid a lump sum for the gang, and after he splits it there is no per worker record left
- Advances are handed over in cash, written on a loose sheet, and remembered differently by the two people involved
- The month closes with somebody typing the whole register into Excel, which is where the real errors enter
- An inspector asks for the attendance record and the only evidence is one person’s handwriting
What replaces the notebook
Hazri tied to the site
Draw a fence around the plot and set a radius. A check-in is only accepted inside it, so the day is attached to a place rather than to whoever was holding the pen.
A face on every day
Check-in takes a selfie. The record is tied to a person, not a name in a column, which is what makes it usable a month later when the day is disputed.
Works with no signal
Punches are captured on the phone and queued with their real timestamps, then synced when the phone finds a network. Basements and remote sites are the normal case here, not an exception.
One worker, several sites
A worker moved between sites mid week checks in wherever they are standing. Nobody picks a site from a list, and the hours land against the right site on their own.
Wages out of the same record
Daily or monthly rates, overtime, advances and deductions, payslips, and PF and ESI ready reports, all calculated from the attendance you already marked. Nothing is retyped.
A month that closes itself
The total is there on the 30th because it was built up day by day. The week of typing at month end is the cost the register hides, and it is the one this removes.
The register is not the problem. The 30th is.
A hazri register works. It needs no charge, no signal, no training and no subscription, it survives a site with no electricity, and a contractor who has run one for fifteen years is not doing something primitive. Any page that opens by calling the notebook stupid is talking to somebody who has never stood on a site at 7 in the morning. So the argument here is not about the notebook.
The argument is about three moments where the notebook has nothing to offer, and all three are moments where money or trust is at stake.
The first is payday. A worker says 26 days, the book says 24. The book is not evidence of anything except what somebody wrote down, and the two days in question were three weeks ago. Whoever is more insistent usually wins, which means either the worker is short or you are paying for days nobody worked. Neither outcome is one you chose.
The second is the end of the month. Every entry in the book has to be counted, half-days resolved, overtime added, advances subtracted, and the result typed into a sheet. That is days of work, and it is work done under time pressure by a person who is copying numbers, which is exactly the condition under which numbers get copied wrong. The errors that reach a payslip are almost never errors of attendance. They are errors of transcription.
The third is an inspection, or a claim, or a case. Somebody official asks for the attendance record and the record is a book in one person’s handwriting, filled in at whatever time suited them, with no way to show that any particular person was on any particular site on any particular day. That is not a comfortable position to be in and it is entirely normal.
What an app changes is that each day carries its own evidence as it is created: a time, a position inside the site fence, and a photograph. Not because anyone assumes your supervisor is dishonest, but because a record that can prove itself removes the argument before it starts, and the argument is the expensive part.
Register, Excel sheet, app: what each one actually gives you
| Paper hazri register | Excel sheet | Shiftelio | |
|---|---|---|---|
| Marking the day | Supervisor writes names, at whatever time they get to it | Typed up later from the register or from memory | Each worker or the supervisor marks in and out at the moment it happens |
| Proof the worker was on site | None. It is handwriting | None. It is a cell in a sheet | GPS position inside the site fence, plus a selfie at check-in |
| Works with no mobile signal | Yes | Yes | Yes |
| Survives rain, dust and a lost bag | No | Only if somebody backed the file up | Yes |
| Half-days and overtime | Written as a fraction, argued about on payday | A formula somebody has to maintain | Calculated from the in and out times, visible all month |
| A worker on two sites in one week | Two registers, reconciled by hand | Two tabs, reconciled by hand | One record, hours attributed to each site automatically |
| Time to close the month | Days of typing | Hours, plus the typing that fed it | The total is already there on the 30th |
| Turns into wages, PF and ESI | No | Only with a second sheet somebody built | Yes |
| Editable after the fact without a trace | Yes, and nobody can tell | Yes, and nobody can tell | Edits are recorded, with who and when |
| Needs a phone on site | No | No | Yes. One per worker, or one supervisor phone for the gang |
Where this is the wrong product
This needs a phone on site. If there is none, buy a biometric terminal instead.
There are exactly two ways to run this. Every worker installs the app and marks their own hazri, or one supervisor keeps a single phone and marks the whole gang in a round, taking a selfie of each worker as they go. The second covers most labour sites, including gangs where nobody owns a smartphone, because it needs one device and not forty.
What it does not cover is a site where nobody will carry a phone at all, or where no supervisor will do the round because the gate rush is 200 people in ten minutes. For that, the right purchase is a fingerprint or face recognition terminal mounted at the gate: a one time cost of roughly Rs 8,000 to Rs 25,000 depending on the model, no per worker phone, and a queue that moves in about a second per person. Buy that. It is a better fit than anything on this page, and we would rather say so now than have you find out in month two.
A terminal has its own trade: it is fixed to one gate, so it tells you somebody entered the building and not which of your four sites they spent the day at, and it does nothing at all for a gang working out of a van. Sites that are spread out are where the phone wins, and that is the honest shape of the choice.
The other apps on this shelf, and when to pick one of them
Laborbook and Superlabor
Both are built tightly around the contractor use case: a labour register, a wage sheet, an advance ledger, often with a Hindi interface and a very short learning curve. If what you want is a digital version of the notebook and nothing more, they are a fair choice and they are cheap. What you are trading away is the verification layer: an entry made by a supervisor in an app is still an entry made by a supervisor, and without a fence and a face attached to it, it settles a payday argument no better than the book did.
StaffApp and AttendFirst
Attendance-first products in the same price region, usually with GPS in some form. The questions worth asking before you buy either are the three that decide whether an app survives a real site: does a check-in work with no signal and sync later, is there one record when a worker moves between two sites, and does the attendance actually turn into a wage calculation or does somebody still export a sheet. Ask them on a trial with your own gang rather than reading a feature list. A trial week costs you nothing and answers all three.
Jibble
Genuinely good software with a free tier that is not a trick, and if your need is time tracking for a small team it may be all you need. It is built for a global market rather than an Indian site, which shows in the places you would expect: Indian payroll, PF and ESI, and the advance and lump-sum arrangements that a labour site runs on are not what it is for. Free is a real advantage. Retyping hours into a separate payroll sheet every month is a real cost, and it is the one that grows with the gang.
Where we differ, in one line
The check-in is verified rather than asserted, the record survives a site with no network, one worker can span several sites without being entered twice, and the wage run comes out of the same record with PF and ESI ready reports. The app reads in 9 languages including Hindi, Marathi, Bengali, Gujarati, Tamil, Telugu, Kannada and Punjabi, which matters when the gang is migrant and the supervisor is not.
Putting it on a live site without stopping work
- Draw the siteDrop a pin on the site and set a radius. A plot gets a wide one, a shop counter a tight one. Punches outside it are not accepted, so a hazri cannot be marked from a tea stall two kilometres away.
- Add the workersName, phone number, daily or monthly rate, and which sites they can work at. Bulk upload from a spreadsheet if you already have the list. A worker can belong to several sites from day one.
- Decide who holds the phoneEither every worker installs the app and marks their own hazri, or one supervisor marks the gang in a round from a single device. Mixed gangs work too: the ones with phones mark themselves and the supervisor covers the rest.
- Run one week alongside the registerKeep writing the notebook for a week and compare. This is the step people skip and the one that settles the argument on site, because the two records disagree on exactly the days that were being disputed before.
- Close the month from the appWages, overtime, advances already deducted, payslips, and PF and ESI reports come out of the same attendance you have been marking all month. Nothing is retyped, which is where the errors used to come from.
The lump sum, the advance, and the cash nobody wrote down
The common arrangement is that a mate or thekedar is paid for the gang and splits it himself. That is a legitimate way to run a site and nothing here asks you to stop. What it normally destroys is the per-worker record: once the lump sum is split, there is no document showing what any individual was owed or received, so a worker who says he was paid short cannot be answered and neither can the person who paid him.
So attendance stays per worker regardless of how payment is arranged. Each person has their own days, each day has its own time, position and photo, and the amount owed against those days is calculated per person even when the money leaves as one payment. You keep the arrangement and you keep the evidence.
Advances work the same way. A cash advance is recorded against the worker with the date and the amount, and it is deducted from the wages automatically rather than remembered. Where an advance goes to a shared house rather than to one person, it is split across the residents to the paisa and deducted from each of them, which is the case that a loose sheet of paper has never once got right. Every advance and every repayment carries an audit trail, so the question of what is still outstanding has one answer instead of two.
What it costs, per year, for the whole site
| Plan | Per year | Covers | Sites |
|---|---|---|---|
| Basic | Rs 999 | up to 2 staff, starts with 3 months free | 1 location |
| Growth | Rs 5,999 | up to 25 employees | 3 locations |
| Business | Rs 14,999 | up to 100 staff, custom branding, priority support | unlimited locations |
| Enterprise | Custom | 100+ staff, API access, dedicated support, SLA | unlimited locations |
What that works out to per worker
A team of 25 on the Growth plan is Rs 5,999 a year, about Rs 240 per person for the whole year, which is less than a day of one person’s time spent typing a register into Excel. A contractor running 40 to 100 workers across any number of sites is Rs 14,999 a year on the Business plan. A two person operation is Rs 999 a year and starts with 3 months free.
Labour attendance: the questions contractors actually ask
What is a labour attendance app and how is it different from a hazri register?
It is an app that records who was present on site, when, and where, instead of a supervisor writing names in a notebook. The difference is not neatness. A register entry is one person’s handwriting made at whatever time they got round to it, and it cannot prove anybody was on site. An app record is stamped with the GPS position, the time, and a selfie taken at the moment of check-in, so the day carries its own evidence. At month end the register has to be typed into Excel by hand and the app has already added the hours up.
Does every worker need a smartphone?
No, but somebody on site needs one. There are two ways to run it. Either each worker installs the app on their own phone and marks their own hazri, or a supervisor keeps one phone and marks the whole gang from it in a kiosk-style round, taking a selfie of each worker as they go. The second is what most labour sites actually use. If nobody on site will carry a phone at all, this is the wrong product and a fingerprint or face terminal at the gate is the right one. We would rather tell you that now than after you have paid.
Will it work on site where there is no mobile signal?
Yes. Check-in and check-out are captured on the phone with the GPS fix and the photo held locally, then queued. When the phone reaches signal, at the gate, in the van, or back at the office, the queued punches sync in the order they happened and keep their original timestamps. Basements, tunnels, plant rooms and remote sites are the normal case on a labour site, not the exception, which is why this is in every plan rather than an upgrade.
How does it stop a supervisor marking hazri for people who never came?
Three things together, and no single one of them is enough on its own. The check-in is only accepted inside the geofence you drew around the site, so it cannot be marked from home. It carries a selfie, so the record is tied to a face rather than a name. And mock-location apps are detected, so a spoofed GPS position does not pass as a real one. What you get is not a system nobody can ever cheat. It is a system where cheating leaves a trail somebody has to explain.
Can one worker be marked at two different sites in the same week?
Yes, and this is the case a register handles worst. Attach a worker to as many sites as you like and they check in at whichever one they are standing in. The punch is matched to the site by position, so nobody picks a site from a dropdown and nobody picks the wrong one. Hours are attributed to each site as they happen, so you can answer what a particular site cost you in labour without splitting a total by hand.
How are half-days and overtime handled?
They are calculated from the in and out times rather than argued about. You set the standard day and the overtime rule once, and the app applies it to every worker every day. A short day shows as the hours actually worked, and overtime shows as the hours beyond the standard, both on the same screen the worker can see during the month rather than for the first time on payday. Most payday disputes on a labour site are about a specific day three weeks ago that neither side can now reconstruct, and a per-day record with a time and a photo ends that conversation quickly.
We pay the mate or thekedar a lump sum and he splits it. Does that still work?
Yes, and you keep both halves of the record. Attendance is still marked per worker, so you know exactly how many days each person did, and you still hold the evidence for each of those days. Advances and cash payments are recorded against the individual with an audit trail, and a group advance paid to a shared house can be split across the residents to the paisa and deducted from each person’s wages. What that gives you is the thing a lump-sum arrangement normally destroys: a per-worker record you can produce when somebody says they were paid short.
Can I use this as the attendance record for a labour inspection or a wage dispute?
It is designed to be the record you produce, and it is a stronger one than a notebook, because every day carries a timestamp, a GPS position and a photo rather than one person’s handwriting. Attendance registers and wage registers are still your legal obligation under the applicable labour law and rules, and this page is not legal advice: check the format your state and your contract require, and export from the app into that format. What the app removes is the situation where the only evidence you have is a book that has been written up from memory.
Does it also do the wage calculation, or only attendance?
It does the wages, and this is the reason to move off the register at all. The same verified attendance feeds payroll directly: daily or monthly wage rates, overtime, advances and loan deductions, payslips, and PF and ESI ready reports. Nobody retypes the register into Excel, which is where the errors normally enter. If you only want attendance you can use only attendance, but the two being one record is what saves the week at month end.
What does it cost for a contractor with 40 workers across three sites?
Rs 14,999 a year on the Business plan, which covers up to 100 staff and unlimited sites. There is no per worker charge, so 40 workers and 95 workers cost the same, and GPS is not an add-on. A team of 25 or fewer on up to 3 sites is Rs 5,999 a year, about Rs 240 per person for the year. A two person operation is Rs 999 a year and starts with 3 months free. All figures exclude 18 percent GST and were verified against our pricing page in August 2026.
Read next
Best attendance apps for construction in India ·Offline attendance ·Multi-site attendance ·Attendance without a smartphone
Get started free