Swizzle
2026Product concept and design, full CMS build
The tool I would have wanted as a bartender and manager: a curated place to concentrate cocktail knowledge and make it easy to use.
Where it started
I spent years managing bars and training bartenders before moving into software. One thing I carried with me was the classics list we used for training: the drinks I expected someone behind the bar to know and be able to make well. Building something around that knowledge had been in the back of my mind since I became a developer.
There are already enormous cocktail databases online, but volume isn't necessarily useful. Good recipes can get buried among thousands of variations, and even the definition of a "classic" can be loose.
What I wanted to build
I imagined Swizzle as the tool I would have wanted as a bartender and manager: a curated place to concentrate that knowledge and make it easy to use.
The library grew from two main sources: the classics list we used at ECC and training material attributed to Sam Ross and Milk & Honey. Where provenance is known, I cite it; where cocktail history is disputed, I don't pretend otherwise.
How I built it
Rather than treating each drink as an article, I structured methods, families, ingredients, glassware, garnish and other characteristics as data. It took more work upfront, but gave search, collections and menus a common foundation.
The scope also grew quickly: personal collections, custom menus, technical guides, community recipes, team tools and back-bar planning. Instead of waiting to build all of it, I focused the first release on the reference library and initial personalization features.
Where it led
Swizzle launched publicly in 2026 and is now gathering usage, feedback and suggestions. What began as a way to share knowledge from my previous career has become a product I can keep building around how people actually use it.
The next decisions are less about what I can add and more about what users find useful enough to build next.