Platform Knowledge Graph API Dataset Developers Documentation Enterprise Request API Access
Home/Knowledge Graph

The Restaurant
Knowledge Graph

Millions of restaurants, hundreds of millions of dishes, and the billions of relationships between them — one living, queryable web of food.

Restaurant Menu Dish Ingredient Nutrition
Entity model

Typed entities, connected by meaning

Every node in the graph has a type, rich attributes, and edges to everything it is related to.

Restaurant

Identity, location, cuisine, hours, price band and brand lineage. The anchor entity every food application starts from.

Menu & Dish

Sections, items, modifiers, descriptions and prices. Resolved across languages and formats into canonical dishes.

Ingredient & Allergen

The hidden layer. Every dish knows what it contains — ingredients, allergens, dietary labels and substitutions.

Cuisine

A cultural taxonomy connecting dishes to their cuisines, regions and traditions — so searches understand meaning, not just words.

Nutrition

Calories, macros and health labels computed at dish level and traced back to ingredients, with confidence scores.

Price

Per-location, per-time pricing intelligence. Track how a bowl of ramen moves across cities, seasons and delivery channels.

Live traversal

Follow one path through the graph

GRAPH TRAVERSAL · Tokyo Ramen Shop
Tokyo Ramen Shop            // Restaurant  · 水
  └─ Menu (2026)             // Menu  · 木
      └─ Tonkotsu Ramen        // Dish  · 火
          ├─ Pork Bone Broth     // Ingredient  · 土
          ├─ Chashu              // Ingredient  · 土
          ├─ 720 kcal             // Nutrition  · 金
          └─ Japanese Cuisine     // Cuisine  · 木

// Cross edges resolve instantly:
Tonkotsu Ramen  4,182 restaurants · 31 countries
Pork Bone Broth  9 cuisines · 214 dishes
"contains egg"  allergens propagate to 1.2M dishes

Knowledge Graph questions

How are dishes resolved across languages?
Entity resolution combines multilingual embeddings, ingredient signatures and human-verified ground truth. "Spicy Pork Ramen", "Karai Tonkotsu" and "辣豚骨拉面" converge on one canonical dish with confidence scores on every edge.
How fresh is the graph?
Millions of menu changes are ingested daily. Most high-traffic restaurants refresh within hours; the global graph has a median freshness of under 24 hours, with webhooks when your tracked menus change.
Can I query relationships, not just entities?
Yes. The graph API supports path queries — "all pork-free Japanese dishes under $15 within 3km" — alongside the standard entity endpoints. See the API docs for graph traversal syntax.

Query the world's food graph.

Get API access and start traversing restaurants, dishes and ingredients in minutes.