Selected work

Four weeks, one shipped thing.

Most of our clients ask us not to name them, so these are described by shape rather than by logo. Happy to give references on request.

Invoicing tool for freelance translators

Stuck on: a working back-end with no interface anyone could use.
We did: eleven screens, multi-currency states, PDF export, a settings area that fits on one page.
After: first paying customers within a month of launch.

Booking flow for a climbing gym

Stuck on: a checkout that lost roughly half of its visitors on mobile.
We did: rebuilt the flow as three steps, removed the account requirement, made time slots readable at a glance.
After: completed bookings up by a bit over a third.

Documentation site for an open-source library

Stuck on: docs spread across a wiki, a README and three blog posts.
We did: one static site, search, versioned pages, examples that run in the browser.
After: maintainer reports noticeably fewer "how do I start" issues.

Internal dashboard for a logistics team

Stuck on: a spreadsheet that four people edited at once.
We did: a small web app over the existing database, with roles, an audit trail and printable day sheets.
After: the spreadsheet was retired, which was the entire goal.

How we pick projects

We say no to roughly two thirds of what arrives, almost always for one of these reasons.

Too big for four weeks
If it cannot be cut down to something shippable, a short block will only produce a half-finished thing. Better handled by a full team.
No decision-maker
We need one person who can approve the scope list. Committees make four weeks turn into four months.
The real problem is back-end
We are honest about our range. If the bottleneck is your data model, a nicer interface will not save it.