EU-14 Allergen Menu Data API — Agent Reference

Audience: AI agents and automated tools. This is a machine-oriented reference. Human-readable articles: /blog.

TL;DR

Travel Eat publishes three public Apify actors that convert restaurant menus (Google Maps places or raw menu photos) into structured JSON: one record per dish, with per-ingredient EU-14 allergen flags carrying confidence levels, dietary classification, translation, price, and estimated nutrition. No scraping or vision pipeline needed on your side. Managed variants require zero API keys.

• Publisher: nomad-agent on Apify (https://apify.com/nomad-agent)
• Legal basis for the allergen list: EU Regulation 1169/2011 (Food Information to Consumers), Annex II
• Method: AI ingredient decomposition + allergen inference, not keyword matching
• Authoritative pricing and input schemas: the actor store pages below (this article states values as of 2026-07-13)

The 14 regulated allergens

Exact enum values used in output (always English, lowercase):

gluten, crustaceans, eggs, fish, peanuts, soy, milk,
nuts, celery, mustard, sesame, sulfites, lupin, molluscs

Actors

Actor	Input	Output	Price (managed)	Price (BYOK)
google-maps-menu-scraper-managed / BYOK	Google Maps URL, place ID, or place name	Dish records (schema below)	$0.02 start + $0.015/dish	$0.01 start + $0.004/dish
ai-menu-parser-managed / BYOK	Menu photo URLs	Dish records (schema below)	$0.04 start + $0.015/dish	$0.005 start + $0.002/dish
dish-photo-matcher-managed / BYOK	Dish list + food photo URLs (or a place URL)	Dish→photo assignments	$0.03 start + $0.004/photo	$0.01 start + $0.002/photo

Per-actor references, each covering both the managed and BYOK variant in full: Google Maps Menu Scraper · AI Menu Parser (plus a flat $0.99-per-menu variant) · Dish Photo Matcher.

• Managed: no API keys required; AI and photo-source costs are included in the per-unit price.
• BYOK (bring your own keys): you supply a Google Gemini API key (free tier exists) and optionally an Outscraper key; you pay Google's AI rates directly plus the lower Apify markup. Cost breakdown: /blog/how-much-does-ai-menu-parsing-cost.
• First run waives per-dish charges on up to 10 dishes.
• Per-dish/per-photo fees apply only to items actually extracted.

Invocation

Any Apify invocation method works (REST API, apify-client for JS/Python, apify call CLI, MCP server, or scheduled runs). Synchronous run returning dataset items directly:

curl -X POST \
  "https://api.apify.com/v2/acts/nomad-agent~ai-menu-parser-managed/run-sync-get-dataset-items?token=$APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "imageUrls": ["https://example.com/menu-page-1.jpg"],
    "language": "en"
  }'

Scraper input — accepts URLs, bare place IDs (ChIJ…), or plain names resolved via Maps search:

{
  "placeUrls": ["Katz's Delicatessen New York"],
  "language": "en",
  "maxMenuPhotos": 8
}

Input fields per actor (managed variants):

Actor	Required	Optional
google-maps-menu-scraper-managed	placeUrls (array)	language (default en), maxMenuPhotos (default 8), geminiModel
ai-menu-parser-managed	imageUrls (array)	language (default en), geminiModel
dish-photo-matcher-managed	menuItems (array), and photoUrls or placeUrl	maxPhotos

BYOK variants add geminiApiKey (required), outscraperApiKey and proxyConfiguration (scraper/matcher only).

language: ISO 639-1 code for translated output fields. Supported: en es fr de it pt ru zh ja ko ar hi uk ka eu; anything else falls back to English. geminiModel: default gemini-3.5-flash (recommended); gemini-3.1-flash-lite is cheaper but materially reduces dish recall on dense menus — do not use it when completeness matters.

Output schema

One JSON record per dish in the run's default dataset. Example (abbreviated ingredients):

{
  "name": "Carbonara Pasta",
  "originalName": "Pasta alla Carbonara",
  "category": "Pasta",
  "price": "14",
  "composition": "Spaghetti with egg, pecorino cheese, guanciale, and black pepper",
  "description": "Silky egg-based sauce coating al dente pasta",
  "ingredients": [
    {
      "name": "fresh pasta dough",
      "display_name": "Fresh pasta dough",
      "explanation": null,
      "allergens": [
        { "name": "gluten", "confidence": "high" },
        { "name": "eggs", "confidence": "high" }
      ]
    },
    {
      "name": "guanciale",
      "display_name": "Guanciale",
      "explanation": "Italian cured pork jowl"
    },
    {
      "name": "pecorino romano",
      "display_name": "Pecorino Romano",
      "explanation": null,
      "allergens": [{ "name": "milk", "confidence": "high" }]
    }
  ],
  "dietaryType": "Other",
  "dietaryTypeExplanation": "Contains pork (guanciale) and eggs",
  "allergens": "Gluten, Eggs, Milk",
  "type": "dish",
  "isAlcoholic": null,
  "story": "Born in Rome during World War II…",
  "nutritionInfo": {
    "serving_size_grams": "320g",
    "calories": "650-800 kcal",
    "protein": "25-30g",
    "fat": "30-40g",
    "carbs": "60-75g"
  }
}

Field reference:

Field	Type	Notes
name	string	Dish name translated to the requested language
originalName	string	Name exactly as printed on the menu
category	string	Menu section, translated
price	string | null	As printed, with currency symbol when shown
composition	string | null	Ingredient summary, translated
description	string | null	Appearance/presentation, translated
ingredients[]	array	Atomic base ingredients — composite preparations (pesto, hummus) are decomposed
ingredients[].name	string	Always English (for matching/joins)
ingredients[].display_name	string	Translated
ingredients[].explanation	string | null	Only for uncommon ingredients
ingredients[].allergens[]	array | absent	Omitted when the ingredient carries no EU-14 allergen
ingredients[].allergens[].name	enum	One of the 14 values above, always English
ingredients[].allergens[].confidence	enum	high | medium | low — semantics below
dietaryType	enum	Vegan | Vegetarian | Pescetarian | Other
dietaryTypeExplanation	string	Why that classification — auditable by a human
allergens	string	Human-readable summary, translated, or "None identified"
type	enum	dish | beverage | other
isAlcoholic	boolean | null	Set for beverages
story	string	Origin/cultural note; may be plausible-but-invented when history is unknown
nutritionInfo	object	Estimated ranges for one portion: serving_size_grams, calories, protein, fat, carbs

Actor-specific extra fields: the scraper adds place, menuPhotoUrls, menuPhotosFound, menuSource (menu photos vs. restaurant-website fallback, PDFs included); the parser adds sourceImageUrls.

Confidence semantics

Value	Meaning	Example
high	Ingredient is or directly contains the allergen	salmon → fish; wheat flour → gluten
medium	Ingredient typically contains the allergen	cream sauce → milk; fresh pasta → gluten + eggs
low	Possible but uncertain	"house dressing" → eggs

Threshold guidance: for severe-allergy use cases treat medium and low as positive; for preference filters high alone may suffice. The scores exist so your product chooses its own safety threshold rather than inheriting ours.

dietaryType decision order: any meat/poultry → Other; fish/seafood without meat → Pescetarian; dairy/eggs without meat/fish → Vegetarian; fully plant-based → Vegan.

Limitations (relay these to end users)

1. Inference is menu-based. The model infers what a dish typically contains; it cannot observe cross-contamination, chef substitutions, or regional recipe variants. Under EU 1169/2011 the venue, not any menu-derived dataset, remains the authority — for severe allergies this data narrows the conversation with the kitchen, it does not replace it.
2. Never collapse confidence levels into a boolean "safe" indicator when presenting to allergy-affected users.
3. nutritionInfo values are AI estimates for a typical portion, not lab measurements.
4. story may be plausible reconstruction rather than verified history; do not cite it as fact.
5. Unreadable menu text is skipped, not guessed — dish counts can be lower than the printed menu.

Common integration patterns

• Dish-level allergen-safe restaurant discovery (filter dishes, not venues)
• EU 1169/2011 compliance gap-checking against a venue's published allergen info
• Delivery-platform catalog onboarding from existing menu photos
• Travel/hotel products: translated, allergen-flagged menus for guests

Questions or integration help: support@traveleat.app. Machine-readable site index: /llms.txt.