It's all about the mouthfeel.

MOUEEL puts two ingredients next to each other and tells you what happens in the mouth. It is not a recipe site: no quantities, no method. Only what holds together, how far off the beaten track that pairing sits, and where it lands on sour, sweet, savoury, bitter and umami.

A chef uses it to find the row they had not thought of. A home cook uses it to find out why the thing in the fridge wants the other thing in the fridge. It is the same table for both.

This is a beta, so some of it is wrong. Found something? Tell me :)

MOUEEL is cooked by Bruno Alves (@pillbet) as an initiative of Platelier.

Somebody should stand behind a tool that tells you what to cook. So: one person wrote the pairings, chose the tiers, and is answerable for the ones that turn out to be wrong.

Four books are the spine: The Flavor Bible, The Flavor Thesaurus, The Flavor Matrix and The Art & Science of Foodpairing. Around them sit recipes from all over, classic cuisines through to Michelin-awarded kitchens, read one plate at a time.

The reason attached to each pairing is written, not generated. That is the slow part, and it is the part worth paying for.

Every pairing is ranked by how far off the beaten track it is, not by a percentage.

classic

It has a name and a country. Tomato and basil.

common

You have eaten it a hundred times without noticing.

uncommon

A cook nods. A civilian raises an eyebrow.

rare

One cuisine does it, or it lives on tasting menus.

wild

Aroma theory says yes. Tradition has never tried it.

Affinity measures precedent, not quality: each joint scores where the catalogue files it. Classic 100, common 88, uncommon 72, rare 58, wild 45, unrecorded lower… and a set scores the average of its joints, so one we have written a lot about scores higher than one we have not.

So a low score is not a verdict on your dinner. Chicken and tomato scores badly here, and butter chicken exists.

Five hues, one job each: sour, sweet, savoury, bitter, umami. Tier is carried by form instead, in dot rails, glyphs and borders. It costs a little legibility, and it buys a palette that can only ever say one thing. That is the point.

• The wild tier is aroma theory, not precedent. François Benzi, a perfumer at Firmenich, put jasmine next to pork liver in Sicily in 1992 because both carry indole. Heston Blumenthal got to white chocolate and caviar the same way, and the shared trimethylamine was confirmed afterwards. Checked at scale in the 2010s, shared compounds turned out not to predict which pairings anyone likes. So a wild row is a mechanism, not a recommendation, and classic and common do not rest on it.
• Holes in the table. Where no pairing is filed, the app says so: a hole in our pairing table, not in the cooking.
• Molecules where they are doing the work. Estragole in tarragon, trimethylamine in caviar. Where the catalogue names a family instead, nobody here could pin it down.

There is no account. What you save lives in your own browser, and any combination is a link. The address bar is the save button.

Page views are counted, and that is all. Cloudflare Web Analytics sets no cookie, stores no identifier and cannot follow you anywhere else. It tells us which pages get opened and how fast they loaded, and nothing about who opened them. So we still cannot tell whether a pairing was any good, and we still need you to say. Nothing is sold, and there is nothing here to leak.