Anyone can promise. Here's what we did.

Hacked servers, bugs that hid for years, a backend deleted the night before a demo. Real cases, real numbers.

When it was on fire

Illustration: a garage turned into a small shop, a flood of parcels pouring in through the open door, inside a man and a woman calmly working at a new packing table with a label printer and a conveyor, parcels lined up in neat yellow stacks

Two weeks after launch, hundreds of orders a day

Two weeks after launch, a small husband-and-wife webshop was flooded with hundreds of orders a day, and customers started leaving angry comments. In a few days we put things in order: support, automatic invoicing, shipping, stock and a CRM.

2 weeks after launch, the peak arrivedRead the case

Illustration: two hands lift a small shop off a cracked, crumbling base with a broken padlock and smoke next to it, and carry it to a new, solid base with a yellow top

RescueDevelopment3 min read

From a hacked server to safety in 12 hours

The server of a webshop with over a billion forints in yearly sales was hacked after five years without maintenance. We moved it somewhere safe in 5 hours, and we've been rebuilding it alongside live traffic ever since.

5 hours to the moveRead the case

Illustration: night with a moon and stars on the left, a rising sun on the right; hands quickly assemble a building from white blocks, from an empty dashed outline to a finished building with a yellow top

RescueDevelopment2 min read

From zero at 8 pm to a live demo at noon

Just before launch, a partner lost the entire backend of a bilingual video job interview system, backups included. The demo was the next day at noon, with around EUR 90,000 of EU funding at stake. We rebuilt it from scratch overnight.

16 hours from the call to the demoRead the case

Illustration: on the left, a cracked box full of small houses with smoke rising from it; on the right, the same houses in neat rows, each on its own pedestal with a yellow shield; two hands carry the last one across

RescueOperations2 min read

A hundred hacked sites, fixed over one weekend

On a Friday, an agency partner found that around a hundred of its WordPress sites had been hacked, along with every available backup. Over the weekend we moved all hundred to a new server and restored them properly. Not a single site has been hacked since.

~100 sites restored over one weekendRead the case

When money was quietly leaking

Illustration: yellow coins travel through a pipe towards a vault and fall out through a hidden crack; one hand finds the crack with a magnifying glass, another patches it with a yellow block

Audit2 min read

The bug nobody knew about

A webshop with over a billion forints in yearly sales was provably losing customers to a hidden payment bug for four years. Nobody knew, not even the client. We spotted it. The damage: nearly half a billion forints. The fix took a few hours.

~HUF 500M lost over four yearsRead the case

Illustration: scattered papers being swept up on the left, a neat stack of paper in the middle next to columns of parcels and coins joined by dotted lines; a hand places a yellow card with a check mark on top of the stack

Audit2 min read

16,000 invoices, one by one

A webshop issued an invoice for every order, but nobody tracked what happened to the parcel or the money. We matched everything line by line and found out what the real revenue was.

16,000+ invoices reconciledRead the case

When something new had to exist

Illustration: a small yellow vintage car speeding along with its hood open, while two mechanics on a flatbed lower a big yellow racing engine into it mid-drive

Development3 min read

From MS-DOS to an API. Under 0.3 seconds. Without flying to China.

At a garment factory of one of our Chinese partners, an MS-DOS machine is the heart of an old production line, and nobody may touch it. It still had to be connected to the CRM through an API, with responses under 0.3 seconds, without stopping the production line. We did it, entirely remotely.

<0.3 s response time from an MS-DOS systemRead the case

Illustration: a flood of heavy framed photographs pours into a yellow press machine and flies out the other side as light image cards towards phones and laptops

DevelopmentScaling3 min read

20 million images a day. Smaller, faster, and nobody waits.

We built an image optimisation API backend for an international company, and we have run it ever since. Scaled to 20 million images a day, it compresses, converts, resizes, crops around the subject, and writes alt text with AI.

20 million images a dayRead the case

Illustration: a messy pile of paper pours onto a conveyor belt on the left, a big yellow machine processes it in the middle, and a single neat stack of identical pages comes out on the right with a yellow check mark on top

MFOIDevelopment2 min read

Tens of thousands of EU documents, in Hungarian in minutes

MFOI needed tens of thousands of EU grant and funding documents translated into Hungarian, in precise legal and financial language. We built an AI translation system trained for exactly this: nearly 100 times more cost-effective than a translation agency, and minutes instead of weeks.

~100× more cost-effective than an agencyRead the case

Illustration: a small truck at a barrier gate, a pole-mounted camera scanning it with a yellow beam, the barrier lifting on its own next to a yellow check mark; the guard booth is empty and a crossed-out access card lies on the ground

Development2 min read

No card, no guard: 1.2 seconds at the gate

A municipality serving over 400,000 residents ran access to dozens of sites with key cards, manual barriers and security guards. We built AI-based vehicle and personnel recognition: it decides in 1.2 seconds, with no access cards and no on-site guards.

1.2 s from camera to barrierRead the case

Illustration: a stream of blank picture frames flies from the left into a big yellow lens, and comes out on the right sorted into four yellow trays, each showing a different simple shape: a circle, a triangle, a square and a face

DevelopmentScaling2 min read

Millions of images a day, recognised in under a second

A growing enterprise's image recognition service buckled at peak times. We built a horizontally scaling API: it processes millions of images a day with sub-second response times, and handles 10x traffic spikes without manual intervention.

10× traffic spikes, no manual interventionRead the case

For us, IT isn't work, it's our passion. It's what gets us up in the morning.

Which of these sounds like yours?

How can we help? We reply within one business day, and in an emergency we step in within 5 hours.

Let's talk