coresmith.dev
A team, seven years, twenty projects

About

coresmith.dev is an independent software engineering studio. Design, code, deployment and the part afterwards where it has to keep running — all of it in-house, by the same team. There is no account manager standing between you and the people building the software, and nothing is subcontracted without telling you.

The team is deliberately small and the work is deliberately narrow. That is a real constraint rather than a positioning statement — we take fewer projects at once, so a start date is a genuine limit and not a negotiating position.

How this works

Scope first, price second, code third

Most projects go wrong before anyone writes a line, in the gap between what was asked for and what was needed. So the first conversation is about what has to change in your business, not about which framework to use. That conversation is free and usually takes one call.

Then you get a fixed price for an agreed scope, or a monthly retainer if the work is genuinely open-ended. No price list, because a price quoted before scope is a guess dressed as a number.

The work is delivered end to end: design, build, deploy, monitor. At handoff the code is in your repository under your account, with documentation and a runbook, and you are not locked into us to keep your own software alive.

What this is built on

Four things we will not trade

It has to run in production

Three of the four systems on this site are running in production today. The fourth is an open experiment and says so on its own page. A demo that impresses and never ships is not work, it is a portfolio piece.

The numbers are measured

Every figure on this site was measured rather than estimated: load times from a real browser against the live sites, page weights from the network, route counts from the sitemaps. If we cannot measure it, we do not claim it.

You own what you paid for

Code in your repository, documentation you can hand to someone else, a runbook for the boring parts. The test is simple: could another engineer take this over without calling us. If not, it is not finished.

The honest answer, early

If the thing you asked for is the wrong thing, you hear it before you commit rather than three months in. Sometimes the honest answer is that you do not need a project at all.

How we work

Six habits that decide the outcome

Understand the task before choosing the tool

Web app, desktop app, integration — those are already answers. The question they answer has to be asked first, or you get a well-built solution to the wrong problem.

Fewer parts, working correctly

Every dependency, every service and every screen is something that can break at three in the morning. The smallest system that does the job is usually the one still standing in two years.

Measure instead of arguing

Disagreements about performance, structure or priority get settled with a measurement. It takes an hour and ends the discussion permanently.

Weekly demos you can click

Not status reports. Something running that you can open and try, every week, so a wrong direction costs a week rather than a quarter.

Written scope, in plain language

What it does, what counts as done, and explicitly what is not included. The last part prevents more disputes than the first two combined.

Delivery includes the day after

CI/CD, monitoring and the runbook are part of the work, not an upsell. A system nobody can deploy or debug was not delivered, it was abandoned in place.

Where this stands

The numbers, plainly

7years
20projects
6countries

Start with the scope conversation. It costs nothing and it is usually one call.