Web application development
Angular and React over .NET or Node, deployed into your own cloud tenant. The systems we build carry school records, safety certificates and order books, so they are judged on uptime and audit trails rather than on animation.
Book a 30-minute scoping call See what we have shipped
If a website is what you need, we will say so and point you at people who build those.
| Who signs in Roles, tenants, single sign-on | Day 1 |
|---|---|
| What it connects to ERP, CMMS, payment, legacy database | Day 1 |
| What must never be lost Records, certificates, audit history | Day 1 |
| Where it runs Your tenant, your region | Day 1 |
| What replaces what Migration from the current system | Day 1 |
What we are good for
Our work sits where a web application holds data somebody is accountable for. That is a different job from publishing pages.
Use a website builder when
WordPress or Webflow will do this faster and cheaper than we can.
Build an application when
This is our work. Every build below replaced something that had stopped coping.
Plain definitions
Neither is a technicality. Both change the budget by a factor, so they are worth settling on the first call.
A web application runs in a browser but behaves like software rather than like a document. It holds state, it enforces who can do what, and it changes something in a system of record when somebody clicks.
A useful test. If two people with different job titles open the same URL and see different things they are allowed to act on, you have an application. If they see the same page, you have a website.
Most of what reaches us is modernisation rather than a first build. An application that already runs the business, written for conditions that have changed, with nobody left who wrote it.
Power Apps, Retool and similar tools are genuinely good for an internal form over a database that one team uses. We say so when that is the answer, because building from scratch would cost you more for less.
They run out of room in predictable places. Per-user licensing that scales badly, integrations with legacy systems that expose no modern interface, data volumes that make the generated queries crawl, and any obligation to hold records where a platform vendor cannot reach them.
What we build
What people ask us for
Almost every enquiry we get is one of these, or two of them at once.
Our work
Case study, Norway
The client ran a system that could no longer handle its data volumes or meet current security standards, and a previous vendor had missed the delivery date. We rebuilt it in modules and put a separate team on migration alone, so schools kept working while years of sensitive records moved across.
Case study, Germany
A cloud-native software firm came to us with poor code quality and missed dates. We staffed a lead and senior engineers first, stabilised the build, then scaled. React and Node on AWS, with GraphQL between the layers and Selenium coverage from the start.
How the work runs
Week 1
Roles, integrations, data, and what is being replaced. You leave with an architecture sketch and a cost band.
Week 2 to 4
Data model, screens, and a clickable prototype before production code.
Week 5 onwards
A demo every two weeks. Priorities change between sprints, not through change requests.
Go-live
Migration rehearsed before the real run. Repositories, pipelines and documentation are yours.
Technology and integration
Every item here comes from something we shipped. It is not a list of what we could learn.
Who does the work
Applications like these outlive the people who commission them. So we keep the team small enough that the engineer who designed your data model is still around when you want to change it.
You interview whoever we put forward. Two of our largest engagements started as recoveries of somebody else's code, which is a decent test of whether a team can read as well as write.
How we work
The models differ in one thing: how much of the management you hand over. Everything else, including who owns the code, is the same in all five.
Changing model later is normal. Augmentation into a dedicated team is the common direction, and the people stay.
Engineers join your team and work inside your process, your board and your code review.
A team that works only on your product, with a lead on our side running delivery.
A long-term engineering site under your standards, where we carry recruitment, HR, payroll and equipment.
04
One price for a scope that is genuinely settled, with the overrun carried by us.
Everything in the centre model, with the whole team moving into your own Vietnamese entity on an agreed date.
FAQs
We ship both. Angular where the team already knows it or the app is form-heavy and long-lived, React where the interface is more fluid. If you have an in-house team, we use what they can maintain after we leave.
Yes, and two of our largest engagements began that way. We read the code first and tell you what is salvageable before quoting.
Yes, as its own workstream with row counts and checksums on both sides. We moved years of sensitive school records with nothing lost.
Your own Azure or AWS tenant, in the region your policy requires. You keep the pipelines.
Your company, entirely, in every engagement model.
Usually 3 to 6 months for a focused application, longer for anything replacing a core system. The scoping week gives you a real number for your build.
Integrations move the number more than screens do. A focused application that replaces one process is usually 10 to 14 weeks. A system the business runs on, with roles, integrations and reporting, runs longer. How it is billed follows the engagement model you pick, which is set out above, and the scoping week ends with a figure for your build rather than a range.
Usually yes, and it is normally the safer route. We rebuilt the Norwegian school platform module by module while schools carried on using the old one, with a separate team on migration alone.
Roles and tenant boundaries are enforced server side, so nothing depends on a hidden button. Identity runs through Azure AD B2C or Keycloak, data sits in your tenant in the region you require, and on the Norwegian platform the security workstream covered SPT, SSAT and NVA testing.
The one that only two people understand, or the one a vendor left half finished. Thirty minutes and you will know what replacing it involves.
