Internal tools and customer portals
The systems a company runs on: ordering, scheduling, inventory, claims, approvals. Usually replacing a spreadsheet that four people are afraid of, and usually worth more than anything customer-facing.
We build web and mobile software for companies that already have customers and a problem worth solving in code. Three projects at a time, no more, and no subcontracting.
A studio of fourteen cannot be good at everything. What is not below, we will name two firms for and take nothing for the introduction.
The systems a company runs on: ordering, scheduling, inventory, claims, approvals. Usually replacing a spreadsheet that four people are afraid of, and usually worth more than anything customer-facing.
React Native unless there is a real reason not to, and we will tell you when there is. Shipped to both stores by us, with the release process handed over and documented.
Accounting, payments, logistics, POS, government APIs. Unglamorous, frequently the highest-return work we do, and the thing most often quoted at half what it costs.
The single biggest cause of a failed software project is a fixed price agreed before anyone understood the problem. So we sell the understanding first, cheaply, and you can take it elsewhere.
Ninety minutes. What the problem is, who has it, and whether software is even the answer. About one in six of these ends with us saying it is not.
Two engineers and a designer, ten days. You get a written specification, a clickable prototype, a risk list and an estimate at ±10%. All of it is yours whatever you do next.
Something running at the end of every second week, on a real URL, that you can use. Not a status report — a build.
Your repository, your cloud account, your CI. We commit into your infrastructure from day one so that "handover" is a meeting, not a migration.
A retainer if you want one, and a clean exit if you do not. We have never held a client's deployment keys hostage and we never will.
Before discovery we quote a range of about ±30%, because that is the honest width of what anyone can know. Studios that quote a single number before understanding the problem are either padding it or planning a change request.
Your GitHub organisation, your cloud account, your domain. We are collaborators on your repository, not the other way round. It makes leaving us easy, which is the point.
Fourteen people cannot properly staff five. When we are full we say so and give you a date, rather than starting you and then thinning the team out in month two.
Everyone who writes code on your project is on our payroll and in our office. It costs us more per day and it is the reason the day rate is what it is.
Full case studies, including the one that ran over, are on the work page.
Replaced a shared spreadsheet three people maintained by hand. The win was not the algorithm — it was that the data stopped being retyped four times.
96 engineer-days · web app + integrationA distributor's resellers were ordering through a marketplace and paying its commission. Now they order through an app the distributor owns.
148 engineer-days · React Native + backendEleven disconnected tools consolidated into two. Most of the work was migration and reconciliation, not features, and we said so before starting.
210 engineer-days · ran 6 weeks overWe pick the technology your next developer is most likely to already know, not the one that is most interesting to us. It is your codebase after we leave.
It gives you a range and tells you what widens it. If the number is far from what you had in mind, that is worth knowing in four minutes rather than after three meetings — and it is the same range our engineers would quote you.