← All
AI-assistedAI-assisted web app · Leiden, The Netherlands

Leiden Home & Hound

For Leiden Home & Hound, the goal wasn’t just to build a pet-sitting website — it was to earn the trust of people handing over the keys to their home and the care of an animal they love.

Leiden Home & HoundLeiden Home & Hound — desktop
Leiden Home & HoundLeiden Home & Hound — mobile
LocationLeiden, The Netherlands
Built with
AIGraphic designNode.js SQLiteManaged hostingMaintenance
The project

What we built

This custom-built platform was designed around what sets Sherry and John apart: professional, nurse-level care for elderly and medically fragile pets — medications, injections, dressing changes — delivered with the warmth of the tagline “Loving Care When You Aren’t There.” Rather than a generic services page, the design leads with calm and credibility: a restrained sage-and-ink palette, the owners’ black line-art logo, self-hosted typography, and paw-print details that keep things personal without tipping into cute.

The site balances reassurance with utility.

A rolling availability calendar shows standard, peak, and booked days at a glance; a clear rate-and-policy page sets expectations before the first conversation; and a booking-request form routes straight to the owners’ inbox with an automatic acknowledgement to the client. A detailed pet-care intake form works both online — arriving as a formatted PDF — and as a printable A4 version, so every client’s routines, vet contacts, and medication schedules are captured once and kept close at hand.

Behind the scenes, a simple admin lets the owners mark peak seasons, block dates, update prices and policies, and manage requests without touching code. The result is a quiet, trustworthy online presence that does exactly what a small, careful business needs: makes it easy to say yes.

How we built it

AI wrote most of the code. Thirty years decided what to write.

The difference from a "vibe-coded" app is everything after step one.

01

Spec first, with AI as a sparring partner

We wrote the product spec and data model with AI, pressure-testing edge cases (what counts as a closure? what about partial hours? where does the data come from?) before a line of code existed.

02

AI-assisted build, engineer-reviewed

Features generated with AI coding tools, then read, tested and corrected by us. Every commit reviewed by a human who has shipped production software since 1997.

03

Data pipelines with guardrails

Closure data comes from multiple sources in multiple languages. AI helps normalise it; validation, logging and fallbacks make sure a bad feed never becomes bad advice on the water. Report back to admins if an API fails.

04

Deployed on our managed EU hosting

Monitoring, backups and patching on the same hosting we offer clients. Built in weeks, not months, and still ours to run and improve.

The honest split

What AI did. What we did.

AI

  • Generated the bulk of the application code
  • Wrote first-pass tests and documentation
  • Produced UI variants to choose from

Wicked Code

  • Decided what to build and what to leave out
  • Reviewed, corrected and hardened every piece
  • Designed the data validation and failure modes
  • Hosting, monitoring, and answering the phone
Want one of these?

We'll build your product the same way.

AI-assisted from the spec onward, reviewed and hosted by people who've done this for three decades. Tell us the problem; we'll tell you honestly whether AI belongs in the answer.

Scope a build

Have an idea that should exist by spring?

Let's talk