rentonhead

Gastromancy

An AI kitchen assistant: photograph your fridge and get a real recipe — shaped by your mood, diet and allergies, and read aloud while you cook.

Client
Own product
Year
2026
Role
Product design · Mobile engineering · Prompt design
Status
Preparing for release

What’s in the fridge becomes what’s for dinner.

Problem

Recipe apps start from recipes. People start from whatever is left in the fridge — and from their diet, allergies and energy level that evening.

Approach

A React Native app built on Gemini 2.5 Flash with structured JSON output. Vision recognises ingredients from a photo of the fridge or counter; mood modes — lazy day, chef, student budget — change the style of the recipe; a saved diet and allergy profile is enforced in every request; and a hands-free voice mode reads each step aloud. Recipe cards can be saved to the photo library.

Result

A complete app with subscriptions and a published privacy policy in three languages, being prepared for the App Store.

What’s inside

  • Fridge photo → ingredients → recipe
  • Mood modes: lazy day, chef, student budget
  • Diet and allergy profile respected in every answer
  • Hands-free voice steps
  • Shareable recipe cards

Stack

React NativeExpoGemini 2.5 FlashVisionSpeechRevenueCat

Under the hood

The code layer of this project — modules and moving parts.

gastromancy/tree --depth 2
app           React Native · Expo → iOS & Android
ai            Gemini 2.5 Flash · vision · structured JSON output
modes         lazy day · chef · student budget
profile       diets · allergies · excluded ingredients (on device)
voice         step-by-step text-to-speech, hands-free
billing       RevenueCat · monthly & yearly with 7-day trial