01 / 10

A till for the informal economy.

Mozambique's shops, bars and mercearias run on memory and paper. kaixa gives them a till, the stock behind it and the day's result - free, forever, for everyone.

Ten screens: the problem, the product, the moment it exists for, how a merchant gets in, why it costs nothing, what protects it, how it reaches people, what comes next, and exactly where it stands today.

Open the app The prototype build

Internal - prepared for Tiago and Hugo Morgado, 3 August 2026.

A merchant's own till, minutes after signing up

02 / 10

The problem

A quarter of a million businesses that keep no books.

250,000+: Handling customers - bars, bottle stores, mercearias, quiosques. Verified against the customer mirror, 20 July 2026.
10-15,000: transactions a day between Handling and those customers.
85%: of Mozambique's workforce is unbanked. No record means no credit and no visibility - for them or for anyone selling to them.

They are not disorganised. They are un-instrumented.A trader knows every selling price by heart. What they cannot tell you is yesterday's takings, which staff member sold what, or whether the fridge is short - because nothing has ever written it down.
The tools that exist were built for somebody else.Point-of-sale software assumes a card machine, a laptop, a bank account and English. This assumes a cheap Android and a person who has never used an app like it.

03 / 10

What it is

A till, and the brain behind the till.

Caixa - the register.Tap the product, tap Cobrar. Counter mode or tables. Cash, card, M-Pesa, eMola on every sale.
Gestao - everything underneath.Stock, expenses, the day's close, reports, products, staff PINs.
Every number ties to an append-only log.A void is a new event, never an edit. Nothing can be quietly rewritten - which is what makes it a control, not a notebook.

"This isn't an operating system. This is a simple tool to help people run the day to day of their business."

Hugo, 3 August 2026

04 / 10

The moment it exists for

Fecho do Dia.

At the end of the day the owner taps once and sees what a pen-and-paper trader has never been able to see without help.

What you sold today.And what it cost you, so the number at the bottom is profit rather than turnover.
Where the money should be.Cash box against card against M-Pesa against eMola - the count that catches the difference between a bad day and a leak.
What should still be on the shelf.Product by product. Go and look; the gap is the answer.

If that screen makes a merchant ask when they can have it, the product works. Everything else in this deck is in service of getting somebody to it.

05 / 10

Getting in

From a phone number to your own shop in six taps.

Five questions, and not one of them is hard.Your number, the code we text you, what you sell, what it is called, where it is. Selling price is the only number a trader carries in their head, so nothing in the flow asks for another one.
The shelf is already stocked when you arrive.Say “bar” and twelve products land priced and ready. Say bar and mercearia and you get both, with nothing duplicated.
It is measured, not asserted.6 taps, 5 questions, 3 fields typed. A build that raises any of those numbers fails the check and has to say what it bought.

Onboarding outranks every other screen in this product. A user who leaves in the first two minutes never reaches Fecho do Dia, so polish spent on experienced users is polish spent on the people who already got past the part that loses everybody else.

06 / 10

The price

Free for everyone. Forever.

What everyone else does.Loyverse - the closest thing to a competitor - gives the till away, then charges for the staff pack, again for detailed inventory, and again to let you keep more than three months of your own sales history.
What we do instead.Nothing is charged, no tier is priced, and nothing in the data model anticipates billing. Handling customers get a VIP tier where deliveries fill the stock in automatically - extra function, never a different price.
Free is the strategy, not a promotion.A merchant who believes the price can change later keeps a paper book as insurance, and a merchant who keeps a paper book never depends on this. Trust buys adoption speed, and adoption speed is the whole game.

"No one's going to feel like they're being held ransom because the switching cost is too high. But that is actually the case."

Tiago, 3 August 2026

07 / 10

What protects it

The moat is everything a merchant would have to leave behind.

Switching cost, earned rather than imposed.By the end of the first month it holds their catalogue, their prices, their staff, and every day they have closed. Nobody re-types that.
Network effects.One platform the whole informal trade is on is worth more to each merchant than the same product used alone - and it is the rail anything sold to them later runs on.
Data that cannot be bought.Complete visibility of a business, including the part not bought from us. VivaCredit already targets 280,000+ informal merchants with credit scored on behaviour instead of collateral; this is the instrument that produces the behaviour.
It is not a Handling feature.kaixa is supplier-agnostic and open to any Mozambican number - bananas or beer. That is what lets it reach shops Handling does not serve, which is where the data gets interesting.

08 / 10

Route to market

The customers already exist, and somebody already visits them.

A distribution network, not a marketing budget.More than 250,000 customers, visited by a salesforce that already carries handheld technology into their shops every week.
A standalone app on the Play Store.Published under the corporate entity, deliberately outside the Handling stable, so it can be sold into sectors Handling has nothing to do with.
Support that scales without people.AI chat and generated instructional video, then a YouTube channel. Assisted onboarding for the largest customers runs as a parallel experiment - never as the requirement.

09 / 10

What comes next

The till first. Then the money. Then the credit.

Now - make it real.A backend, durable sync, a real one-time password, and the delivery weld that fills a Handling customer's stock automatically. Then a pilot.
Next - mobile money.Hugo is deepening the Vodacom relationship on the Handling side. The part that matters here: a kaixa merchant becoming an M-Pesa agent from inside this app.
Then - credit.Behaviour on a merchant nobody else can see, on the one platform they use every day. That is the payoff the free product is buying.
Never - a fee.Every one of the above is worth more with the whole market on the platform than it could ever be worth as a subscription.

10 / 10

Where it stands

Fourteen rounds in. Live, and honest about what is not real.

Real today

The till, counter and tables, with four payment methods
Day close, reopen, voids - all append-only
Stock count, expenses, reports, products, staff PINs
Registration and onboarding for a brand-new shop
Installs to the home screen and works offline
84 automated tests; a design audit at 0 fail, 0 warn

Simulated, and knowingly so

The one-time password - the demo code is 1234
The Handling delivery feed
All data. There is no backend yet; everything lives on the phone
Nobody has used it in a real shop

$5: a month to host, today. Tiago's read: marginal cost per user is very small and falling - manageable even at a million.
0: dependencies, build steps and network requests. It cannot break because somebody else's library changed.
3: open owner decisions: the home domain, when the pilot starts, and whether the green brand goes into the main build.

Register a shop and walk it The brand

Customer figures are the portfolio's approved claims, live-verified 20 July 2026 - re-verify before any external use. Onboarding counts are measured on every build. Quotations are from the 3 August call transcript, not its summary.