UK Menu Data API for Meal Recommendations — Restaurant, Menu and Allergen Data Across Delivery Apps and Restaurant Sites
How Food Data Scrape scoped structured, matchable menu records for a UK taste-intelligence start-up.
Key takeaways
Key takeaways
A UK menu data API gives a recommendation engine one consistent record per dish: restaurant, cuisine, menu section, item name, description, price and any published dietary or allergen tags. For this start-up, Food Data Scrape proposed collecting public menus from delivery platforms and restaurant websites, normalising them into one schema and refreshing them on an agreed cadence.
Client overview
What did the client need?
The client, a London-based start-up, recommends dishes by matching diners' flavour profiles with what restaurants serve, so it needs menus as structured data, not web pages or PDFs.
They asked for restaurant listings (name, location, cuisine), full menu structures, items with descriptions, prices and categories, dietary or allergen tags where available, and an optional refresh. The first phase covered London restaurants across all four source types, with other major UK cities planned as the next phase.
The challenge
What made this hard?
- Different menu layouts. Each platform groups sections, options and add-ons differently.
- Duplicate restaurants. One venue can appear on several platforms under slightly different names.
- Thin descriptions. Many items have short or missing descriptions, which limits flavour matching.
- Inconsistent allergen data. Tags are optional and worded differently; some menus only say "ask staff".
- Channel price differences. App prices can differ from in-restaurant prices, so each price keeps its source.
- Frequent menu changes. Seasonal dishes and sold-out items mean records go stale without a refresh plan.
The solution
How did Food Data Scrape solve it?
We proposed a scoped, source-aware pipeline built around a single menu schema and delivered through an API or scheduled files.
Approach
- Public sources only. Collect publicly visible listing and menu pages from Deliveroo, Uber Eats, Just Eat and restaurant websites, respecting robots.txt, terms of service and rate limits. No logins and no personal data.
- One menu schema. Map every source to the same restaurant → section → item hierarchy so the engine reads one format. See our restaurant menu data scraping service.
- Restaurant matching. Link venues across platforms by name, address and postcode.
- Tags kept as published. Store dietary labels and allergen text exactly as shown, mapped to the 14 allergens in Food Standards Agency guidance, and flag them as unverified.
- Change tracking. Record first-seen and last-seen dates so new, changed and removed dishes stand out.
- Flexible delivery. Offer access through the Food Data Scraping API or JSON/CSV feeds.
Data captured
Data attributes / fields
| Field | Description |
|---|---|
| restaurant_id | Stable internal ID that links one venue across sources |
| restaurant_name | Restaurant name as listed |
| address_postcode | Public business address and postcode |
| cuisine_type | Cuisine labels shown on the listing (e.g. Thai, Italian) |
| menu_section | Menu category or section heading (e.g. Starters, Mains) |
| item_name | Dish or drink name |
| item_description | Published description text, used for flavour matching |
| price_gbp | Listed price in pounds sterling, with its source |
| dietary_tags | Published labels such as vegan, vegetarian, gluten-free |
| allergen_notes | Allergen text as published, mapped to the 14 UK allergens where stated |
| source_platform | Deliveroo, Uber Eats, Just Eat or restaurant website |
| last_seen_at | Timestamp of the latest crawl that found the item |
Results
What were the results?
- A model-ready menu format. The scoped schema gives the recommendation engine one consistent structure for every restaurant.
- Clear source trail. Each price and tag carries its platform and last-seen date.
- Safer allergen handling. Tags are labelled as published by the restaurant or platform, never as safety guarantees, and diners are directed to confirm with the restaurant.
- Refresh-ready records. Change tracking supports updates at the agreed cadence.
Client testimonial
In the client's words
"Having UK restaurant menus in one schema – sections, dishes, descriptions, prices and allergen notes as published – gives our recommendation engine a consistent base to match flavour profiles against. The source and last-seen date on every record also make it easier to keep recommendations current as menus change."
— Head of Product, UK Food-Tech Start-up
Questions
Frequently Asked Questions
Restaurant listings, menu sections, item names, descriptions, prices in GBP, published dietary and allergen tags, the source platform and last-seen dates.
No. They are shown exactly as the restaurant or platform published them. Apps should tell diners to confirm allergens with the restaurant before ordering.
Natasha's Law covers prepacked for direct sale (PPDS) food. Non-prepacked restaurant food follows separate FSA allergen information rules.
Refresh frequency is agreed per project, from one-off datasets to scheduled updates. Track changes with last_seen_at. See our nutrition and allergen compliance data.
Need a UK menu data API?
Tell us your cities, source types and refresh cadence, and we will scope a free sample. Visit fooddatascrape.com to discuss your data requirement.
