Sabotage
Site Store and Editor

- Role
- Front-end development
- Design
- CCT
- Stack
- Next.js, Sanity, Stripe
- Built
- August 2026, 24 days
Impact
~3 seconds from Publish to the live site.
A new site for a London creative studio. I built it alone from the design file in 24 days, and the studio edits every word and picture on it themselves.
- 24 days first commit to last, one developer
- 100+ fields the studio edits on its own



Studio
The studio can change everything on its site without a developer.




Pages
Every page holds the designers' drawings at any size, with one breakpoint.




Looking back
It was really fun executing a vision, building a system and using a stack I hadn't used before. The whole idea was counterculture, and I learned so much about working on a very quick timeline.
Scope
Every page in the designers' file: the homepage strip, the projects, the store, the Smoking Area, the legal pages and the 404.
The editor behind all of it, set up field by field.
Publishing, which goes live in about three seconds.
The store and the newsletter.
SEO, and a share card for every page.
The guide I handed over with it.
Contributions
How the pages behave at the sizes between the desktop and phone designs. I proposed it to the designers.
Lessons
Measure the file instead of eyeballing it. Every gap and size I measured in the designers' file was one less guess in the build.
The editor is a product too. The studio edits over 100 fields themselves, so each one had to make sense to someone who didn't build the site.
Show what changed instead of explaining it. Each round went out with collages and screen recordings plus a short changelog, so nobody had to hunt for the differences.
Design file to handover
Kickoff: Scope and a week-by-week plan.
Design file: Measured down to the unit.
Build: The site, the editor inside it and the store.
Rounds: Each round came with collages, screen recordings and a short changelog.
Handover: A guide to the editor, and the keys.
Studio, Problem
A studio site goes stale the moment every edit needs its developer, and the people editing this one are designers.
Early on, a change took up to a minute to show. The client refreshed, saw yesterday's strip, and fairly concluded publishing hadn't worked.
Studio, Decision
The editor lives at the site's own address and opens the real page beside the form. Clicking anything on the page jumps to its field.
Every field explains itself in the words the page uses, and soft limits are warnings, not errors: a fifth tag gets "Four is the drawing."
Publishing clears the site's cache straight away, and every page refreshes within 60 seconds anyway if that ever fails.
Studio, Results
Publish to live in about three seconds, and never more than a minute.
Studio, Tradeoffs
A publish re-renders every page on its next visit. For a studio's site that is cheap, and no corner of it can go stale.
Pages, Problem
Every page was drawn at a desktop size and a phone size. Media queries for every width in between would drift away from the drawings.
Pages, Decision
Each kind of page has its own unit, a share of the screen measured against its drawing, and every size is written in those units.
The phone project frame matches the desktop one so closely that project pages need no breakpoint at all. The whole site has one.
Project photographs sit at one height and take the width their own shape asks for, so nothing is cropped.
Pages, Results
Desktop and phone match the drawings with six units and one breakpoint for the whole site.

