The Chilli Index
125+ chilli varieties, heat ratings and grow guides on one data model. Scoped, designed and built by directing Claude Code.
Problem
The first version was WordPress and a pile of plugins. Opening the tracker in a second browser overwrote the data, and about 40 plants lost their records in one go. They are still in the garden with no labels. The real fault was not the plugins: the build had gone straight to building, with no plan underneath it.
What we built
A filterable index of chilli varieties on one clean data model, a side-by-side compare tool for any two varieties, a grow tracker with user accounts to log a season, an events calendar and an ebook. It started with a planning session rather than a build - less than half a day to settle the data model, the user and what good looked like - and every build session afterwards inherited that.
Stack
Next.js and Supabase, deployed on Vercel. Built with Claude Code: Darren has not hand-written any of the code, and oversaw the brand, design system, writing, UX and accessibility himself.
Timeframe
Less than half a day of planning before anything was built, then a couple of months in Darren's spare time. Hand-written, the same product would have taken over a year.
What it does now
Live at thechilliindex.com with 125+ varieties and growing, the compare tool and the grow tracker live. It is the proof piece for the AI Product Sprint: the same process, pointed at a client's product instead of a chilli habit.