What we do
What we can build for you.
Start with what you need rather than what it is called. If none of these is quite it, describe the problem and we will tell you what would solve it.
A website that brings in enquiries
A fast, well-written site that works properly on a phone, gets found on Google, and makes it obvious how to contact you. Yours to edit afterwards, without paying us to change a phone number.
Good if your current site is slow, dated, hard to edit, or does not bring in work.
An online store or booking system
Take orders, payments or appointments online, with the stock, calendar and receipts handled for you. Set up so the money lands in your account and the paperwork looks after itself.
Good if you are taking orders or bookings by phone, email or WhatsApp and losing track of them.
A customer or staff portal
A private, password-protected area where customers check their account, or where your team does its work. Everyone sees what they should see and nothing they should not.
Good if you are emailing spreadsheets, statements or files to customers or staff one at a time.
Custom software for how you actually work
When off-the-shelf software nearly fits but not quite, we build the part that does. Usually that starts small — one process, one screen — and grows as it earns its place.
Good if your business runs on a spreadsheet that only one person fully understands.
Joining up the systems you already pay for
Your website, accounting, CRM and payment tools can pass information to each other automatically instead of someone re-typing it. Fewer mistakes, and hours back every week.
Good if the same information gets typed into two or three different systems.
Fixing or taking over something that already exists
A site that keeps going down, a system nobody left instructions for, or a developer who has moved on. We work out what is there, write it down, stabilise it, and take it forward.
Good if something important is broken, slow, or held together by one person who has left.
For funded startups
Or the specific thing standing in your way.
Teams who can already build usually come to us for one of these — a specialism they are short of, capacity they need for a quarter, or the one control blocking a contract.
First version, in weeks
A working product in front of real users on a fixed scope and a fixed price. We build the one thing that has to be true for the idea to work, and deliberately leave out everything that can wait.
Good if you have a deadline — a demo day, a pilot customer, a board meeting — and a spec that is bigger than the time.
AI features that do not blow the budget
Model features built to a unit cost you have agreed in advance: retrieval that returns the right context, prompt caching where it pays, evaluations so you can tell whether a change made things better, and a cost per request you can put in a pricing model.
Good if a model feature works in a notebook and nobody can say what it will cost at ten thousand users.
Unblocking enterprise sales
The engineering behind SOC 2 and GDPR answers: SSO and SAML, audit logging, role-based access, tenant isolation, data residency, retention and deletion that actually reaches every copy. We build the controls; your auditor signs the report.
Good if a security questionnaire is sitting between you and your largest contract.
Cutting the cloud bill
A measured pass over compute, storage, egress and managed services, with each change costed and its saving verified afterwards. We tell you what is safe to cut, what needs a refactor first, and what to leave alone.
Good if your infrastructure spend is growing faster than your revenue and nobody owns the number.
Getting ready for the load
Capacity modelled from your own numbers, load-tested to find where latency actually turns, then the bottlenecks fixed — queues, connection pools, caching, autoscaling — with SLOs and alerting so you find out before your customers do.
Good if a launch, a campaign or a new contract is about to multiply your traffic.
A data and telemetry pipeline
Device, sensor or event data collected reliably, validated before it is stored, downsampled on a policy you chose deliberately, and available to query without a warehouse bill that outgrows the product.
Good if you are collecting more data than you can query, or losing some of it and not knowing which.
Technical due diligence
An independent read on a codebase and team — architecture, security posture, key-person risk, real cost to maintain, and what the roadmap would actually take. Written for investors and boards, in language they can act on.
Good if you are investing, acquiring, or being asked hard questions about someone else’s technology.
Software that handles regulated data
Health, financial and identity data built to the rules from the first commit: PHI kept out of logs and analytics, consent and audit trails, encryption with keys you hold, retention and deletion that reaches every copy — and the evidence an assessor asks for.
Good if a regulator, a hospital, a bank or an insurer has to be satisfied before you can launch.
Marketplace and platform engineering
The parts of a two-sided business that are hard to retrofit: split payments and payouts, escrow and refunds, ratings and trust, dispute handling, and the matching that decides whether both sides actually clear.
Good if you connect buyers and sellers and the money, trust or matching is held together manually.
Backends for devices and fleets
The service behind hardware: provisioning and identity per device, over-the-air updates that can be rolled back, telemetry that survives a bad connection, and a console your operations team can actually run a fleet from.
Good if you ship a physical product and the software behind it grew as an afterthought.
Turning your data into an API product
A public API worth paying for: versioning that does not break integrations, keys and quotas, rate limits that throttle abuse and not your best customer, usage metering that matches what you invoice, SDKs, and documentation people can start from.
Good if customers want programmatic access and your API was built for your own front end.
Security engineering, not security theatre
Vulnerability management triaged by whether a finding is reachable rather than by its score, secrets out of the repository, dependency and container scanning that blocks in CI, threat modelling on the paths that matter, and a response plan someone has actually rehearsed.
Good if your vulnerability backlog only grows, or a questionnaire asked something you could not evidence.
Accessibility that passes a real audit
Keyboard operability, correct roles and states on custom components, contrast that meets the ratios, reflow at 320px, and status messages that announce. Tested with a screen reader, not just an automated scan — which catches about a third of what actually fails.
Good if procurement asked for a VPAT, or you have never tried the product without a mouse.
Making deployment boring
The pipeline behind the four DORA measures: tests that run in minutes, deployments small enough to be uneventful, migrations that roll back, feature flags, environments that match production, and the monitoring that tells you within minutes rather than from a customer email.
Good if releases are events people schedule around, or a rollback means restoring a backup.
A senior engineer on your side of the table
Architecture reviews, hiring and interview support, code review, and the unglamorous decisions that are expensive to get wrong early. Some companies want this for a few days a month; others want it until they hire.
Good if you are the only senior person, or the first engineering hire has not started yet.
How it is built
The engineering behind it.
For anyone comparing suppliers, or who simply wants to know what their money is buying. Skip it if that is not you — nothing here changes what you get.
[01] Websites & apps
The part your customers see, and everything behind it
The screens people use, the rules that decide what happens when they click, and the database that remembers it. Built together by one team, so nothing falls between two suppliers.
- Websites, online stores and booking systems
- Customer portals and staff dashboards
- Connections to the software you already use — payments, accounting, CRM
- TypeScript, Node.js, Express, Middy.js, Python, Go, PHP, Laravel, C#
[02] Hosting & reliability
Somewhere reliable to run it, and someone watching it
Where your software actually lives. Set up so it stays up when you get busy, so a mistake can be undone, and so we know something is wrong before your customers tell you.
- Hosting set up on Amazon or Google Cloud
- Automatic backups, and restores that have been tested
- Docker and Kubernetes
- Alerts when something breaks, and automatic recovery where it is safe
- Updates that go live without taking the site down
[03] Data & automation
Getting your information out of places it is stuck
Most automation stalls for a dull reason: the information is trapped in a system that will not share it. We fix that first, then automate the job.
- Reports that build themselves instead of being assembled by hand
- Information pulled in automatically from other systems on a schedule
- Moving data out of an old system into a new one without losing any of it
- AI features where they help — drafting, summarising, classifying
[04] Specialist work
Two things most studios will turn down
Phone systems and blockchain payments. Both are common asks and hard to hire for, and we have delivered in each.
- Automated phone menus and call routing for contact centres
- Testing that phones the system itself to check it still answers correctly
- Crypto wallet sign-in and payments
- Smart contracts written and reviewed
Common questions
The things people ask before they get in touch.
Including the two most sites avoid: what it costs, and how long it takes.
I do not know exactly what I need. Can you still help?
Yes, and that is the usual starting point. Describe the problem the way you would to a colleague — what is slow, what is manual, what customers complain about. Working out what should be built is part of the job, not something you have to arrive with.
What does a project cost?
It depends on what is being built, so we quote per project rather than publishing a price list. A small fix or a straightforward website is a matter of days; a portal or custom system is measured in weeks. You get a written quote with a fixed scope before any work starts.
How long will it take?
Most first releases are between two and eight weeks depending on size. We deliberately start with the smallest version that is genuinely useful and put it live, rather than disappearing for six months.
I already have a website or system. Can you take it over?
Yes. We take on work other people started, including systems with no documentation and no original developer available. The first step is always to find out what is actually there and write it down.
Who owns what you build?
You do — the code, the accounts, the domain and the hosting, all in your name. If you ever want to move to another supplier we will hand it over cleanly and help them pick it up.
What happens after it goes live?
You can take it in-house, or we can keep looking after it — hosting, updates, backups and monitoring — for a fixed monthly amount. Either way you get written instructions for running it.
Do we need to be in the same country?
No. We work with clients anywhere and everything is done remotely — calls, shared documents and a link you can check progress on whenever you like.
Will I understand what you are telling me?
That is on us. You talk to the person doing the work, not an account manager relaying it, and anything technical gets explained in ordinary language.