I just launched Chefrise on Android.
The original problem was simple: I would look at the ingredients I already had and still not know what to cook.
So I built a small product around that moment.
Chefrise lets you:
- add or scan pantry items
- ask an AI chef for recipe ideas based on your pantry
- save recipes
- turn missing ingredients into a shopping list
- complete recipes and build cooking streaks
- see XP and league progress
The interesting part was not just the AI.
The harder product question was retention.
A recipe app can easily become something people open once, generate a recipe, and forget. I wanted the loop to feel more like Duolingo: small progress, streaks, XP, leagues, and a reason to come back tomorrow.
The stack is Expo React Native, Supabase, Edge Functions, and AI recipe generation. Codex helped me move quickly on implementation, but the product psychology took the most thought.
A few things I learned:
- AI is only useful when it starts from real user context.
- Pantry state matters more than a generic recipe prompt.
- Streaks should reward meaningful action, not just opening the app.
- Mobile UX details like haptics, keyboard behavior, push routing, and fast session restore matter a lot.
- “Generate recipe” is not the product; the habit loop around cooking is the product.
Android is live now:
https://play.google.com/store/apps/details?id=com.chefrise.mobile
The landing page is here:
I’m still improving the loop and would value feedback from people who cook at home or build consumer apps.