Things we build because we want them to exist
Our own products keep us honest. We pay their hosting bills, watch their analytics, answer their users, fix the bugs and work out how to get people to care.
Everything that teaches us comes back into our client work — what things actually cost to run, which AI features justify themselves, what breaks at scale, and what happens in month six rather than week one.
Pingsy
Our productDon’t send a card, send a Pingsy
Pingsy turns a card into a page. Build a personal digital card or page with messages, photos and links, and Pingsy hosts it for you. Share it as a link, or as a QR sticker that can go on a gift, a bottle of wine or a business card.
Why we built it
A greetings card is thrown away within a fortnight. The things people actually want to send — a video, a photo album, a message from everyone at the office — have nowhere to live.
The interesting product problem
Making a page-builder simple enough that someone creates a card in four minutes on a phone, without it becoming so limited that every Pingsy looks identical.
The interesting technical problem
User-generated pages at consumer volume: cheap hosting, abuse handling, and QR stickers whose destination must keep working for years after the sticker is printed.
What it taught us
The economics of user-generated content are unforgiving — storage and moderation costs scale with usage while the price does not. That shapes how we advise clients on anything with an upload button.
- Digital cards
- QR
- Gifting
LaPared
Our productAI match analysis for padel
LaPared turns one fixed camera into a full match analysis. Computer vision tracks the ball, detects players and attributes each shot, so players and coaches get stats that used to need a multi-camera system.
Why we built it
Padel grew faster than its analytics. Professional-grade analysis existed but required installed multi-camera rigs that no community club was going to buy.
The interesting product problem
Amateur players want to know whether they are improving. Expressing that from raw shot data, without turning it into a spreadsheet nobody reads, is the actual product question.
The interesting technical problem
Tracking a small, fast ball through glass walls, occlusion and variable lighting, from a single fixed viewpoint — and processing video at a cost per match that a consumer price can absorb.
What it taught us
Video processing is where theoretical cost models meet reality. Getting per-match compute down to something sustainable taught us more about cloud cost control than any client project has.
- Computer vision
- AI
- Sports tech
TeleStats
Our productFootball knowledge, tested
A football data platform covering more than 36,000 players across 35 years. Play several game modes, or build your own quiz with the community game builder.
Why we built it
We wanted to find out what you can build on top of a genuinely large, messy historical dataset once it is properly structured.
The interesting product problem
A quiz is only fun if the difficulty is right, which means knowing how obscure each of 36,000 players is — a property no dataset ships with.
The interesting technical problem
Reconciling decades of football data where names, clubs and competitions all change over time, then making it searchable fast enough to feel instant.
What it taught us
Data quality work is most of the project and none of the demo. We budget for it honestly on client data projects because of what this one cost us.
- Football
- Data
- Games
Shifted Fitness
Our productTraining that follows your cycle
A fitness app built around the female hormone cycle. Workouts and nutrition adapt week by week to how your body changes, with an AI wellness coach for questions in between.
Why we built it
Mainstream fitness apps apply a model developed on male physiology and leave women to adjust it themselves.
The interesting product problem
Personalisation that adapts weekly has to stay comprehensible. If a user cannot tell why today’s session changed, the adaptation reads as randomness.
The interesting technical problem
Shipping a consumer iOS app: app store review, subscriptions, retention, and an AI coach whose per-user cost has to stay well under the subscription price.
What it taught us
Consumer app economics are brutal and instructive. Acquisition cost, churn and the real price of an AI feature per subscriber are things we now model before recommending them to anyone.
- Mobile app
- Women’s health
- Fitness
We'd build yours the same way
Same discipline, same attention to what it costs to run. See the client work, or tell us what you have in mind.
Start a project →