safepak.org

Our story

Built in Pakistan, for its guests.

How it started

It started with a taxi fare

The moment was specific. A foreign YouTuber, filming his first hours in Pakistan, trying to get from the airport to his hotel. He did not know which apps operate here, so he took a street taxi. He paid roughly double the fair fare, in a car barely holding together, and he said so on camera to an audience of people deciding whether to visit.

None of that was malice. He simply had no way to know what a fair fare was, or that fixed-price apps existed at all. That gap, between what a visitor can find out and what any local already knows without thinking, is the entire reason SafePak exists.

Three attempts

We built it wrong twice before this

The first version was PHP and MySQL. It worked, and we threw it away, because it could not do the one thing that mattered most: function when the network does not. Somebody standing at an airport with no SIM card is exactly who needs us, and a database-driven site gives them a blank page.

The second version over-corrected. A hand-built framework with no dependencies, admin two-factor authentication, cache invalidation, offline architecture. Genuinely good engineering for a product that did not yet have verified information to put inside it. We were building a vault before we owned anything worth locking up.

The version you are reading runs as static files with the emergency page cached on your device, so it opens in airplane mode. That took three attempts to understand, and admitting it is cheaper than pretending we planned it.

The hard parts

None of the hard parts were code

Design was harder than engineering. Our earlier interface was competent and completely forgettable, and it took someone telling us plainly that it looked a decade old before we rebuilt it from the ground up.

Funding is zero. Every service behind this runs on a free tier, chosen deliberately so the product survives whether or not anyone ever writes us a cheque.

But the hardest work by far is research. Confirming which services actually exist, which have quietly shut down, and what things genuinely cost. There is no database to buy. It is phone calls, visits, and reading the fine print, over and over.

What we refused to ship

And what we got wrong in public

The SOS button that did nothing

Our first prototype showed a confident message saying your alert had been sent to the police. It sent nothing. It also labelled guides as background-checked when nobody had checked anything. Both were deleted rather than launched, and that decision became the rule the whole product now runs on.

Two dead apps we kept recommending

For a while this site told arriving visitors to install two ride-hailing apps that no longer operate in Pakistan. Uber left in 2024 and Careem ended its rides in 2025, and we were still recommending them because we had written that content from memory instead of checking. It is a small mistake that could have left someone stranded at an airport at night. We corrected it, and we now verify service listings against current sources rather than assumption.

We publish these because a product that promises honesty should be willing to name its own failures. The rule that came out of them is written down and governs every screen: the Honesty Principle.

Who built it

Four students, no salaries

Adan Khalid

Project lead

Electrical engineering, NUST CEME

Faysal Ali

Web platform

NUST

Anees Haider

Research and collaboration

NUST

Abdullah

Web security

Air University

In three years, SafePak intends to be the single trusted bridge between foreign visitors and Pakistan’s services, its institutions, and its people.