Tonight, from what you already have
Snap a receipt or your fridge — Brillat sees the ingredients and proposes plates you can actually make right now. No grocery run, no second-guessing.
Gastronomy from your fridge.
Snap what's in your fridge — Brillat finds tonight's plate using 200 years of gastronomy and the chemistry of flavor.
“Tell me what you eat, and I'll tell you what you'll love.”— after Brillat-Savarin, 1825
What Brillat is, in one paragraph
Brillat is an iOS AI chef that reads your fridge or grocery receipt and proposes recipes you can make right now — grounded in the chemistry of flavor and the tradition of Jean Anthelme Brillat-Savarin's 1825 Physiology of Taste. It ships with English, Korean, and Japanese as first-class languages, uses photos only after in-app consent, and learns your palate from one tap of feedback after each meal. Made by Codeful in Hanam, Korea. Support: support@vibeland.app.
Why Brillat
Why this combination works
Every suggestion is grounded in flavor chemistry — Maillard browning, umami synergies (chicken's inosinate × soy's glutamate = ×7), thermal safety. We show the why in one line, not a lab report.
Saltier? Less onion? Got it.
Tap one tag after each cook. Brillat adapts the next recipe within minutes — and remembers the trade-offs you care about. Across cuisines, across weeks.
"Use today" before it expires
We surface ingredients about to go off and propose plates that use them first. The savings tally on your profile screen turns abstract food waste into a number you can feel.
Four steps, one minute each
Snap
Photograph your fridge or your last grocery receipt. OCR extracts items in seconds.
Match
Brillat ranks recipes by ingredient coverage, expiry urgency, and umami synergy.
Cook
Step-by-step instructions with chemistry-grounded heat, timing, and substitution cues.
Learn
Your 👍 / 👎 tunes the palate model within minutes. Next recommendation is yours.
Worked example · the chemistry layer
Inosinate × Glutamate = ×7 umami.
Two simple molecules from chicken and soy combine to amplify perceived umami by roughly seven times. Brillat reads this network beneath every classic dish, so you can substitute, adapt, and improvise without losing the plot.
Five axes Brillat learns about you
Each meal you cook updates a five-axis taste vector. Over a few weeks it becomes a fingerprint sharp enough to predict whether you'll love a recipe before you start.
Umami9.0
Salt6.2
Sweet1.2
Fat4.4
Spicy3.8
Warm Citrus
Cream surfaces, amber-to-burnt-orange accents, lime chips for science. The palette is shared 1:1 between the web and the iOS app so the two feel like a single product.
Cream
#FFFBF0
Surface · light
Burnt Orange
#F97316
Accent 2
Amber
#FBBF24
Accent 1
Lime
#65A30D
Science chip
Warm Brown
#2A1810
Ink
From Physiology of Taste to your fridge
- 1755Brillat-Savarin born
French lawyer-gourmand whose taste essays would later found gastronomy.
- 1825The Physiology of Taste
The book reframes eating as a science of pleasure. Still in print 200 years on.
- 1912Maillard reaction described
Louis-Camille Maillard names the browning reaction behind seared steak, bread crust, and roasted coffee.
- 1908Umami isolated
Kikunae Ikeda identifies glutamate as the fifth basic taste — chemistry catches up to the kitchen.
- 2026Brillat ships
An AI chef that reads your fridge and explains the chemistry beneath every plate it suggests.
Quick answers
What is Brillat?
An iOS AI chef that reads your fridge or grocery receipt and proposes tonight’s plate, with the chemistry of flavor explained in one line.
Who is Brillat-Savarin?
Jean Anthelme Brillat-Savarin (1755–1826), French lawyer and gourmand. His 1825 book The Physiology of Taste founded the modern study of gastronomy.
What chemistry does Brillat use?
Maillard browning (≥140°C), caramelization (≥110°C), umami synergy (inosinate × glutamate ≈ ×7), thermal safety thresholds (chicken 75°C, eggs 70°C), and time-temperature curves for pectin and protein structure.
How does palate learning work?
After each cook you tap one tag (Saltier ↑, Bland, Umami ↑, Onion ↓…) and 👍 or 👎. The taste-vector model updates a small delta per axis and re-ranks future recipes — typically by 0.4–0.6 on a 0–10 scale, propagating to next session.
Languages
English (source), Korean, Japanese — first-class. UI, recipes, ingredients, and chemistry notes are all localized.
Privacy
Photos are uploaded only after in-app consent, processed only by a paid AI inference provider, deleted within 24 hours by default. We do not train models on your data and do not include any advertising SDK.
Pricing
Free tier covers inventory, weekly planner, and a daily recipe limit. Pro unlocks unlimited recipes, full palate history, and recipe library export. Apple StoreKit handles payment.
Platform
iOS 26 and later, iPhone-first (camera capture). Brought to you by Codeful, based in Hanam, Republic of Korea.