If you run a kirana shop in Nepal and you are searching for POS billing software, you are probably being sold the wrong thing. Most vendor pages open with “IRD approved” and a demo button. Very few of them tell you the one fact that decides your entire budget: a typical neighbourhood kirana pasal in Nepal is nowhere near the turnover where the Inland Revenue Department actually forces you into CBMS electronic billing. Knowing that before you pick up the phone can be the difference between a Rs 8,000 setup and a Rs 70,000 one.

This guide is written for the shopkeeper, not the software seller. It covers what POS billing software for a kirana shop in Nepal really needs to do, what the IRD does and does not require at your size, what the software genuinely costs once hardware and renewals are counted, and the practical failures — loose daal with no barcode, udhaaro that nobody reconciles, internet that dies during the evening rush — that decide whether the system is still running six months later.

1. Do you actually need POS billing software yet?

This is the section every vendor site skips, because the honest answer sometimes is “not yet”. Let me give you a threshold you can test tonight.

Where software stops being optional is not the bill count itself — it is the invisible leakage. In a shop doing 60–100 bills a day, three things quietly drain money and none of them show up in the cash drawer:

• Price drift. Your cooking-oil rate changed twice in Ashoj but the person at the counter in the evening still charges the old price. Two rupees per litre on 40 litres a week is about Rs 4,200 a year, from one product.
• Silent stockouts. You run out of a fast-mover on a Saturday evening and the customer goes to the shop across the road — and often does not come back for the whole basket, not just the missing item.
• Forgotten udhaaro. Not the big accounts. The Rs 200 and Rs 350 entries someone forgot to write down. Shopkeepers I have worked with consistently underestimate this, and it is usually the single largest number when we finally reconcile.

POS billing software earns its keep by making those three things visible. If your shop is small enough that you personally know every price, every shelf and every debtor, you are the system, and you do not need to buy one yet.

2. The IRD question: PAN, VAT and CBMS, explained honestly

This is the biggest content gap across every competing page on this keyword. They use “IRD approved” as a scare tactic without ever telling you which tier you fall into. Here is the actual structure, as it stands for FY 2082/83.

Tier 1 — PAN-registered only (most kirana shops)

If your annual turnover is under the VAT threshold, you operate on a PAN and issue ordinary sales bills. Under Section 10 of the VAT Act 2052, VAT registration is mandatory when annual turnover exceeds NPR 50 lakh for goods-only businesses and NPR 30 lakh for services or mixed businesses. A kirana shop selling goods sits under the NPR 50 lakh line. Registration must happen within 30 days of crossing the threshold, and certain sectors — liquor, electronics, motor parts and others — must register regardless of turnover, which is worth noting if you also sell liquor or mobile accessories alongside groceries.

What you need: software that prints a clean bill with your shop name, PAN, item lines, quantity, rate and total, and keeps a sales register. You do not need CBMS integration, and you should not pay a premium for it.

Tier 2 — VAT-registered (larger marts and growing shops)

Every registered business must charge VAT at 13%, file a VAT return by the 25th of every Nepali month, and issue invoices in the format prescribed by Rule 17 of the VAT Rules, 2053. The 13% rate has been unchanged since VAT was introduced in 1997 and the Finance Act 2081 did not change it for FY 2082/83.

What you need: correct tax-invoice format, taxable/exempt separation (this matters a lot for a kirana shop, because many basic food items are exempt under Schedule 1), Bikram Sambat dates, an unbroken invoice series, and a monthly VAT summary your accountant can work from.

Tier 3 — CBMS electronic billing (large retail only)

This is where the fear-selling happens. Mandatory e-billing was extended to firms with annual turnover above Rs 250 million (Rs 25 crore) from the FY 2080/81 budget, and that threshold was subsequently lowered to Rs 200 million (Rs 20 crore) by an IRD notice dated April 2026. The policy brings a much larger pool of medium-to-large enterprises under real-time oversight through the Centralized Billing Monitoring System. Lower thresholds apply to specific sectors — for hotels, restaurants and canteens the threshold is reduced to NPR 5 crore.

If you are in Tier 3, or expect to be within a year or two, the approval process itself is a separate project — I have written it up in detail in how to get IRD approval for billing software in Nepal, and you can check which vendors are already listed in the IRD approved billing software list in Nepal before you commit to anyone.

3. The nine features a Nepali kirana shop genuinely needs

Vendor feature lists run to sixty items. For a kirana counter, nine of them decide whether the software survives. The rest are marketing.

1. Billing in under six seconds per item

Measure this at the demo, with a stopwatch. Ask to bill a realistic basket: 1 kg rice (weighed), 2 packets Wai Wai (barcode), 500 ml oil, 250 g daal (weighed), one recharge card. If that basket takes more than 45 seconds end-to-end including printing, it will fail at 7 p.m. on a Friday.

2. Keyboard-first operation

Touchscreens look good in a demo and slow you down in real use. The fastest kirana counters in Nepal run on a cheap keyboard: type three letters, press Enter, type quantity, press Enter. Anything that forces mouse movement between every item costs you seconds per bill.

3. Decimal quantities for weighed goods, whole numbers for packets

This sounds trivial. It is not. Kirana stock is half loose and half packaged. Good software enforces the difference so a cashier cannot bill 1.5 packets of noodles or round 750 g of sugar up to 1 kg.

4. Nepali and English item search

Your staff think in Nepali. If they have to remember that “चिनी” is stored as “Sugar”, search fails during the rush. The software should let you store both names against one item and find it either way. If you keep product lists or price boards in Preeti font, you will need to convert them to Unicode before import — I built a free Preeti to Unicode converter for exactly this kind of cleanup.

5. Udhaaro at the point of billing

Credit must be a payment method on the billing screen, not a separate screen to visit afterwards. If it takes an extra step, nobody does it during a rush, and the ledger is wrong by evening.

6. Mixed payment on a single bill

Rs 500 Fonepay, Rs 200 cash, Rs 150 on khata — on one bill, in one action. This is the most common real-world payment pattern in Nepali retail now and a surprising number of systems still cannot express it.

7. Low-stock alert tied to the supplier and last rate

An alert that just says “low” is useless. An alert that says “Sunflower oil 1L, 8 left, last bought from Shrestha Traders at Rs 168” turns into a purchase order in ten seconds.

8. A day-end report you can read in one minute

Total sales, split by cash / QR / wallet / udhaaro, bill count, and the drawer figure to match. If the day-end report needs interpretation, it will not be read.

9. Your data, exportable, on demand

Ask for a CSV export of items, sales and party balances before you buy. If the vendor hesitates, walk away — you are looking at a system you cannot leave.

4. Why offline-first matters more than cloud in Nepal

Every vendor writes “works offline” as a bullet. Almost none explain what it means, and the difference between two implementations is enormous.

Test it honestly at the demo: unplug the router, then bill three items and print. If the vendor asks you not to do that, you have your answer.

The second question to ask is what happens to Fonepay and wallet reconciliation while offline. QR code transactions in Nepal grew about 92.5 percent in number in FY 2024/25 over the previous year, and QR is now the most widely used electronic payment instrument. Fonepay is accepted at over 13 lakh stores, from small kiranas to large supermarkets. In practice that means a very large share of your bills are now QR. A static counter QR keeps working when your POS is offline — the customer pays the bank, not your software — but the bill must still record “paid by QR” locally so the day-end split reconciles. Confirm this works with no internet at all.

5. The barcode problem nobody warns you about

Here is the most practical gap in every competing article on this topic. They all promise “barcode scanning” as if a kirana shop is a supermarket. It is not.

Walk your own shelves and count. In a typical Nepali kirana pasal, roughly half to two-thirds of what you sell has no usable barcode:

• Loose rice, daal, sugar, chiura, beaten rice, spices — sold by weight from a sack
• Local ghee, achar and oil in unbranded packaging
• Eggs, bread from a local bakery, milk packets with smudged print
• Cigarettes and loose singles
• Vegetables, if you carry them

A scanner only helps with the branded third. For everything else you need one of three things, and you should decide which before buying:

1. Quick keys / favourites grid. Your 30 fastest loose items pinned to number keys or a single screen. This is the cheapest and usually the best answer for a small shop.
2. In-house barcode label printing. You print and stick your own labels on pre-packed bags of rice and daal. Worth it if you pre-pack in fixed sizes; a waste of time if you weigh to order.
3. Weighing scale integration. The scale sends the weight straight into the bill line. Genuinely useful above roughly 100 bills a day, and overkill below it.

6. Udhaaro khata: digitising the credit book properly

Credit is not a flaw in the kirana model — it is the reason your regulars are your regulars. The copy on the counter fails not because it is paper but because it has no ageing and no reminder.

A properly implemented udhaaro module gives you four things paper cannot:

• One running balance per party, with every entry linked to the bill it came from, so “I already paid that” becomes a screen instead of an argument.
• Part payments against older bills — because that is how people actually settle, Rs 500 at a time.
• Ageing buckets — who is 0–30 days, 30–60, and over 90. This is the single most valuable report in a kirana shop and the one most systems either omit or bury.
• A reminder you can actually send — a Viber or SMS message with the balance and the bill reference, generated from the ledger.

The recovery routine that works

1. Every Saturday morning, open the ageing report.
2. Anything over 30 days gets a polite message with the exact figure and the last bill date.
3. Anything over 60 days gets a phone call, not a message.
4. Anything over 90 days stops getting new credit until something is paid — and the software should be able to enforce a credit limit per party so this is not a conversation you have to have at the counter.

That last point matters. If your staff can extend credit without limit, the ledger will be accurate and still useless. A credit ceiling per customer, set by you and enforced by the software, is the feature that converts a digital khata from a record into a control.

7. What POS billing software really costs in Nepal (2026)

Published Nepali pricing, as listed by the vendors themselves at the time of writing:

Software licence is rarely the biggest number. Budget realistically for the whole counter:

8. What the main options actually offer

Rather than rank vendors, here is how the categories differ, so you can place any quote you receive:

9. A realistic 7-day rollout plan

Most vendors promise “setup in a day”. That is true for installation and false for adoption. Here is a schedule that actually works, run in parallel with your existing copy so you never stop trading.

1. Day 1 — the top 100. Enter only your 100 fastest-moving items with correct rates and units. Not 800. One hundred. This covers the large majority of your bills.
2. Day 2 — set units and tax flags. Mark which items are weighed and which are by piece, and which are VAT-exempt if you are registered. Getting this wrong now is painful later.
3. Day 3 — printer and payments. Test the thermal printer with your actual bill layout, stick the counter QR where the customer can reach it, and confirm the bill records the payment method.
4. Day 4 — staff logins and limits. One login per person. Decide who can give a discount, who can edit prices, and who can see purchase cost. Cashiers should not see margin.
5. Day 5 — run parallel. Bill on the software and keep the copy. Compare at closing. Expect gaps; fix the item master, not the process.
6. Day 6 — load the udhaaro openings. Enter current outstanding balances per party as opening figures. Do this once, carefully, with the copy in front of you.
7. Day 7 — first real day-end. Close on software alone. Match the drawer. If it matches within a few rupees, you are live. If not, find the one item whose rate is wrong — it is almost always a rate, not a bug.

Then: the Dashain test

Your real deadline is not go-live, it is your first festival season. Two weeks before Dashain, pull the previous year’s sales by item if you have it, and use the reorder report to stock ahead. A POS that has not been used to plan a festival stock-up has not yet paid you back.

10. Seven mistakes that kill kirana POS rollouts

• ✓Buying CBMS you do not need. Covered above. Match the tier to your turnover.
• ✓Entering the entire catalogue on day one. Enthusiasm runs out around item 300, and a half-entered catalogue makes the stock module untrustworthy immediately.
• ✓Skipping purchase entry. Sales reduce stock; if purchases never increase it, every figure goes negative within a month and people stop believing the screen.
• ✓No UPS. Losing power mid-bill on a machine with no battery corrupts more days than bad internet ever will.
• ✓Letting everyone use the owner login. Then no report can tell you who gave the discount or voided the bill, which is the main reason to have software at all.
• ✓Not asking about the renewal. Year one is quoted, year two is a surprise. Get the AMC or renewal figure in writing before you pay.
• ✓No export test. Ask for a CSV of items, sales and party balances during the trial. If you cannot get your own data out, you do not own it.

Bringing it together

For most kirana shops in Nepal, the right answer in 2026 is unglamorous: an offline-first billing app in the Rs 8,000–15,000 a year range, run on a modest counter PC with a UPS and a decent thermal printer, using a hundred well-maintained items, a counter QR, and a disciplined udhaaro ledger with ageing. That setup beats an expensive ERP that nobody updates, every single time.

Match your compliance tier honestly, test the software with the router unplugged and a bag of loose daal, get the renewal price in writing, and commit to ten minutes of purchase entry every evening. Everything else is a feature list.

If you are also handling invoices outside the counter — wholesale orders, credit customers, supplier bills — you may find our invoice generator and billing software guide useful alongside this one. And if you expect to cross into VAT or CBMS territory in the next year or two, read how to get IRD approval for billing software in Nepal and check the current IRD approved billing software list before you lock into a vendor.

Frequently asked questions