Hosting, deployment, backups, SSL and monitoring — plus migrations to AWS, Google Cloud or Firebase when the current setup has stopped keeping up.
A website is not finished when it launches. It needs somewhere to run, updates applied, backups that have actually been tested, and a person who answers when it goes down at nine on a Saturday.
Most of the emergencies we get called into were avoidable and cheap to avoid: an expired certificate, a disk that filled up, a backup nobody had ever restored from. Maintenance is dull, and it is the difference between a bad hour and a bad month.
Set up on shared hosting, a VPS or a cloud platform, with a deployment path that does not involve dragging files over FTP.
Automated, off the same server, and periodically restored to check they work. A backup nobody has tested is a hope, not a backup.
Certificates renewed before they expire, platform and plugin patches applied, and the obvious hardening most installs never get.
Uptime and error alerts that reach a human, so a problem is not first reported by a customer who could not check out.
Moves to AWS, Google Cloud or Firebase when growth or cost make the current arrangement the wrong one — planned, with a rollback.
Where it runs, what it costs, what is backed up, what is out of date, and what would happen right now if the server disappeared.
Backups, certificates and patches first. This is unglamorous and it removes most of the risk.
Only when there is a reason — cost, performance or reliability. Migrating because a platform is fashionable is a way to buy new problems.
Alerting configured, plus a periodic note on uptime, updates applied and anything that needs a decision.
We would rather say so now than three meetings in.
A monthly plan sized to what is being maintained, with the response time written down. We will not promise a fifteen-minute response and then not staff for it.
Sometimes, and we can tell you within an hour. If the server takes a second before sending anything, hosting is implicated. If the server responds quickly and the page still takes six seconds, no amount of moving host will help — the page is loading too much.
Yes. We start with an audit and tell you what state it is in, including if the honest answer is that it needs rebuilding rather than maintaining.
You get a contact who responds, and the response time depends on the arrangement we agree. We will not promise a fifteen-minute response and then not staff for it.
Describe the problem rather than the solution. We will tell you what it takes — including when the answer is smaller than you expected, or when cloud, hosting & maintenance is not what you need at all.
What is happening in this corner of the industry right now.
Anthropic says an unreleased model raised the lower bound on the Riemann hypothesis after running for about a day and a half across 60 coordinated sub-agents and 650 candidate approaches. Two staff mathematicians confirmed the result and it was formalised in Lean.
Read moreCode in the iOS 27 beta describes Apple Reference Image, a system that would embed provenance data into photos at capture so they can later be verified as camera-made rather than AI-generated. It would ship switched off and verify through Private Cloud Compute.
Read moreThe Galaxy S27 Pro is set to carry a 12MP 3x telephoto while the Ultra drops its 3x entirely in favour of a 50MP 5x periscope, leaving the two hard to compare directly. A more capable ALoP prototype for the Pro appears to have been shelved to protect the Ultra's position.
Read moreWebsites and web applications built front-end to back-end — designed around what a visitor needs to do, and built so the next developer can pick them up.
What this involvesInternal tools, dashboards and integrations built around how your business actually runs — starting with the smallest version that removes the most pain.
What this involvesAndroid apps from a first version that proves the idea to the release handling logins, payments and notifications — with the yearly platform updates that keep it published.
What this involves
