Crisp

Overview

Choose your air fryer model, or describe what you feel like eating, and find recipes with ingredients, quantities and steps written for that model. The web version is live; the mobile apps are in preparation.

Challenge

Most air fryer recipes are written for no particular appliance, while temperature settings, cooking times and capacity differ from model to model, so people have to adapt each recipe themselves. The challenge was to organize recipes around the specific model a person owns, and to be clear about what has and has not been checked on a real appliance.

Solution

Built the web app with Next.js on Vercel, with recipe data organized by air fryer model and a natural-language AI search alongside the model filter. Recipes are drafted with AI from the manufacturers' official manuals, and the site states that they have not yet been cooked on the actual appliances.

Results

• Web version live at cookwithcrisp.app
• Recipes can be browsed for free without registration
• Available in Japanese, English, Simplified Chinese and Traditional Chinese
• iOS and Android apps in preparation

Learnings

Learned how much the structure of the data (which model, which settings) shapes the search experience, and that stating plainly what has not been verified matters as much as the features themselves. Working with Claude Code made it practical to build the web app and prepare the mobile app as one person.