Peregrino presents

Spotify understood music through
450 attributes. Mesa Futura
understands food through
every dimension that matters.

Nutrition. Preparation. Culture. Mood. Cost. Season. Creator.
A knowledge graph where every recipe lives at the intersection of 25 entity types, 55 relationships, and 559K data points — generating infinite combinations that no competitor can replicate.

5,690
Foods mapped
154
Nutrients tracked
524K
Nutrient values
55
Relationship types
MF

Mesa Futura

Sabor Graph — Food Intelligence Ontology

31
Entities
559K
Records
55
Relations

What can you do with a Food Intelligence Graph?

Each use case traverses multiple layers of the ontology — connections that are impossible with a flat recipe database.

Mood × Nutrition

"I'm stressed. What should I cook?"

A user feels anxious after work. Instead of browsing random recipes, the graph finds foods that biochemically reduce cortisol — and wraps them in comfort.

Graph traversal
MoodTag:stressed → pairs with FlavorProfile:reconfortante,cremoso → Nutrient:tryptophan,magnesium → FoodItem:camote,avena,plátano → Recipe matches
Result: "Atole de avena con plátano y canela" — comfort food that's also rich in tryptophan (serotonin precursor) and magnesium (nervous system support). The recipe knows why it works.
MoodTag → FlavorProfile → Nutrient → FoodItem → Recipe
Culture × Season

"It's September — what's the iconic dish?"

A creator wants seasonal content. The graph knows that August-September = chiles en nogada season, tied to Mexican Independence Day.

Graph traversal
SeasonalityTag:sep → FoodItem:chile_poblano,nogada,granada → CulturalTag:independencia → Cuisine:poblana → Recipe matches
Result: Surfaces chiles en nogada recipes tagged as Poblana cuisine, linked to Independence Day, with in-season ingredient alerts. Suggests content angles: "The story behind the tricolor dish."
Season → FoodItem → Culture → Cuisine → Recipe
Substitution × Allergens

"I'm lactose intolerant — adapt this recipe"

A user loves a recipe but can't eat dairy. The graph doesn't just remove cheese — it finds substitutes with similar flavor, texture, AND nutritional profile.

Graph traversal
RecipeIngredient:queso_panela → Allergen:dairy → FoodItem:queso_panela → SUBSTITUTES_FOR → FoodItem:tofu_firme [similar protein, texture:firme, cost:budget] → recalculate NutrientAmount
Result: Suggests tofu firme (similar protein, firm texture) or cashew cheese (closer flavor). Auto-recalculates macros. The recipe stays delicious and nutritionally equivalent.
Ingredient → Allergen → FoodItem sub FoodItem → Nutrients
Budget × Nutrition

"Best nutrition for 200 pesos — this week"

A family needs to eat well on a tight budget. The graph optimizes across cost, nutrition density, and what's actually available this week at the tianguis.

Graph traversal
PricePoint:tianguis,CDMX,budget → FoodItem → NutrientAmount:cost_per_protein → rank by nutrient density/peso → SeasonalityTag:current_month → Recipe matches with those ingredients
Result: "Lentejas con nopal y chile morita" — lentils are the #1 protein-per-peso food, nopales are nearly free at the tianguis, and morita chiles are in season. Total cost: ~$45 MXN for 4 servings.
Price → FoodItem → Nutrients rank Season → Recipe
Creator × Intelligence

"Generate a week of anti-inflammatory meals"

A creator wants to build a meal plan series. The graph doesn't just filter — it generates combinations that satisfy dietary rules, variety constraints, and nutritional targets simultaneously.

Graph traversal
DietaryProfile:antiinflamatoria [rules: no gluten, no dairy, no refined sugar] → FoodItem pool → NutrientAmount:omega3,curcumin,antioxidants → DishCategory variety [no repeat sub_category in 7 days] → generate 21 meals
Result: A 7-day plan: Mon lunch = "Salmón con salsa de mango y aguacate" (omega-3), Tue = "Caldo de lentejas con cúrcuma" (anti-inflammatory spices), Wed = "Bowl de quinoa con verduras rostizadas"... Each meal is unique, nutritionally optimized, and content-ready.
Diet → Foods → Nutrients → Category → Recipes
Preparation × Culture

"Teach me nixtamalización — and what to make with it"

A user discovers the traditional process. The graph traces the full food transformation chain and shows every dish that depends on it.

Graph traversal
CookingMethod:nixtamalizar → TRANSFORMS FoodItem:maíz_seco → FoodItem:nixtamal → CookingMethod:moler → FoodItem:masa → COMPOSED_OF → [tortilla, tamales, sopes, tlacoyos, huaraches...] → Recipes for each
Result: A knowledge tree: maíz → nixtamal → masa → [12 different antojitos]. Each branch links to recipes, equipment (metate vs molino), regional variations (Oaxaca vs Jalisco masa), and the Nahuatl etymology (nextamalli). This is food education, not just recipe search.
Method → Food → Method → Foods → Recipes

Why Food Intelligence matters

Recipe apps are commodity software. The knowledge graph beneath them is the defensible asset.

01

Recipes are the MP3s. The ontology is the genome.

Spotify didn't win because it had songs — everyone had songs. Spotify won because the Music Genome Project decomposed every song into 450 attributes (tempo, mood, instrumentation, era) that enabled discovery, recommendation, and infinite playlists.

Mesa Sana has 1,000+ recipes. So does any cooking blog. What no competitor has is a Food Genome — the ability to decompose every recipe into its nutritional DNA, cultural roots, emotional impact, economic profile, and seasonal relevance. That decomposition is the product. The recipe is just the surface.

02

Creators supply the content. We own the intelligence layer.

This is the Spotify model applied to food:

Creators provide
  • Recipes (the "songs")
  • Photos and videos
  • Personal style and brand
  • Audience and community
Mesa Futura provides
  • Nutritional analysis (automatic, 154 nutrients deep)
  • Cultural classification and tagging
  • Mood and sensory profiling
  • Cost optimization and seasonality
  • Ingredient substitution intelligence
  • Dietary compliance verification
  • Cross-recipe discovery and recommendation
  • Meal plan generation engine

Creators make content accessible. The graph makes it intelligent. Neither works alone. Together, they create a platform nobody can replicate by just hiring more chefs or scraping more recipes.

03

Every recipe uploaded makes the graph smarter.

When a creator adds a new recipe, the system doesn't just store it — it enriches the knowledge graph:

1
Creator uploads "Tacos de cochinita pibil con cebolla morada en escabeche"
2
Ingredient resolution: each ingredient string maps to a FoodItem node with full nutritional data
3
Auto-nutrition: 154 nutrients calculated from ingredient composition. Diet compliance auto-checked (gluten-free? anti-inflammatory?)
4
Cultural tagging: Yucateca cuisine, achiote preparation, linked to Día de Muertos tradition
5
Mood mapping: comfort food, smoky aroma, rich texture → tagged for "nostálgico" and "celebrando"
6
Economics: achiote paste = budget, pork shoulder = mid-range. Total cost estimated by region.

One recipe upload generates ~50 new graph edges. After 10,000 recipes, the graph has 500,000+ connections. That's the network effect — the moat deepens with every creator.

04

The competition can't catch up.

Recipe apps
Store flat documents. "Chicken tacos" is just a title + ingredient list. No understanding of what chicken is, what a taco requires, or why this recipe is comforting.
AI chatbots
Generate plausible recipes from training data. No real nutrition data, no provenance, no cultural accuracy. "Mole with soy sauce" passes their filter.
Nutrition databases
Have the science but zero cultural context, zero mood mapping, zero creator economy. They're spreadsheets, not platforms.
Mesa Futura
All of the above, connected. Science + culture + emotion + economics + creators, in a single traversable knowledge graph. Each dimension reinforces the others. This is the compound advantage.

The Spotify for Food model

How creators, consumers, and the knowledge graph create a flywheel that accelerates over time.

MF

The Flywheel

1
Creators upload recipes Their content, their brand, their audience. Free to upload, like uploading to Spotify.
↓
2
The graph enriches automatically Every recipe gains 7 dimensions of intelligence. Nutrition calculated. Culture tagged. Mood mapped. Cost estimated. Substitutions generated.
↓
3
Consumers discover through intelligence "What can I cook when I'm sad and it's raining and I have 100 pesos?" — answered by graph traversal, not keyword search.
↓
4
Engagement data feeds back Which recipes work for which moods? What substitutions do people actually make? The graph learns from behavior.
↓
5
Better recommendations attract more creators Creators get more reach because the platform understands their food. More creators = more data = smarter graph. Repeat.
$

Revenue layers — just like Spotify

Creator subscriptions
Pro tools: advanced nutrition analysis, meal plan builder, branded content, audience analytics. Creators pay for the intelligence layer.
Consumer premium
Personalized meal plans, mood-based discovery, dietary tracking, family nutrition optimization. Consumers pay for smart recommendations.
Data licensing
Food brands, supermarkets, health insurers — they all want structured food intelligence. The graph is the API. Nobody else has this data.
Cultural partnerships
Tourism boards, culinary schools, government nutrition programs. "The definitive map of Mexican food" has institutional value.
vs

Mesa Sana today vs. Mesa Futura tomorrow

Mesa Sana (today)
  • 1,000 recipes as flat documents
  • 4 macronutrients per recipe
  • Free-text ingredients
  • Manual diet tagging
  • No substitution engine
  • No mood/cultural intelligence
  • Content-first, data-second
Mesa Futura (tomorrow)
  • 1,000+ recipes in a knowledge graph
  • 154 nutrients per ingredient (524K values)
  • Every ingredient linked to FoodItem entity
  • Automatic diet classification via rules
  • AI-powered substitution graph
  • Mood, culture, season, cost intelligence
  • Data-first, content-powered

The ontology is called Sabor Graph — the flavor graph. Because understanding food means understanding everything that gives it sabor.

The Sabor Graph Schema

25 entity types across 7 layers. Click any entity in the 3D graph for details, or browse them all here.

"The value isn't the recipe. It's understanding food the way Spotify understands a song."

Food Identity

5,690 foods with scientific names, biological taxonomy, and cross-references to international databases. Every chile, every quelite, every cut of meat — identified and classified.

Composition

154 nutrients × 5,690 foods = 524K data points. Not just calories — amino acids, fatty acid chains, minerals. Plus allergen mapping and 15 dietary profiles (Keto to Cuaresma).

Preparation

24 cooking methods including nixtamalizar, tatemar, moler en metate — with Nahuatl names. The first ontology that speaks the language of Mexican food, not just translates it.

Cultural Context

10 regional Mexican cuisines (Oaxaqueña, Yucateca, Norteña...) + traditions like Tamalada, Día de Muertos, Quinceaños. Food isn't just nutrition — it's identity.

Mood & Wellness

No competitor has this. Comfort food mapping, mood-based discovery, sensory profiles (crujiente, ahumado, reconfortante). "I'm stressed" → here's what to cook.

Economics

Cost per nutrient density. Tianguis vs supermarket pricing. 12 Mexican micro-seasons (huitlacoche, flor de calabaza, chapulines). Budget optimization that understands Mexico.

Creator & Content

Where Mesa Sana's 1,000+ recipes live — but now every ingredient resolves to the knowledge graph. Every recipe inherits 7 dimensions of intelligence.

The Moat

Anyone can copy a recipe app. Nobody can replicate a knowledge graph with 559,804 connected data points, 25 entity types, and 55 relationship types built on peer-reviewed nutritional science, traditional Mexican food knowledge, and cultural intelligence. This is Mesa Sana's unfair advantage.

Dimensions
View
1-7 Toggle layers L Labels F Flow R Reset Esc Deselect
Peregrino × Mesa Futura