Restaurant Competitor Menu & Price Monitor

Turn ordinary restaurant URLs into a recurring competitor-intelligence feed. The Actor discovers same-site menu pages, builds a normalized baseline, and reports only material menu and pricing decisions on later runs.

Best for

• restaurant groups benchmarking local competitors;
• franchise and hospitality agencies monitoring many locations;
• food distributors watching category and ingredient trends;
• restaurant-data and market-research pipelines.

Paste URLs, not scraper instructions

Input a restaurant homepage or exact menu page. Exact public ordering URLs give the highest success rate. The Actor follows same-site menu links and extracts from:

• Schema.org Menu, MenuSection, MenuItem, and Product JSON-LD;
• public application-state JSON embedded in the page;
• common server-rendered menu markup and microdata;
• JavaScript-rendered ordering pages, including public Toast menus, through an automatic browser fallback.

No restaurant login, access-control bypass, OCR, or AI billing is used. Modern ordering pages commonly reject data-center requests, so the default input uses Apify residential proxy traffic and only launches a browser when the faster HTML path returns no items.

Intelligence events

• ITEM_ADDED and ITEM_REMOVED
• PRICE_UP and PRICE_DOWN
• AVAILABILITY_CHANGED
• DESCRIPTION_CHANGED
• CATEGORY_CHANGED

Every event contains old/current evidence, integer-cent price movement, percentage movement, source URL, severity, extraction confidence, detection time, and a deterministic fingerprint. A run summary ranks the largest price moves and counts changed restaurants.

Noise controls

Use minimumPriceChangePercent, keywords, and eventTypes to deliver only changes relevant to a buyer's watchlist. emitFullSnapshot turns the same Actor into a normalized menu feed; leave it off for an exception-only monitoring workflow.

First and later runs

The first run always writes one BASELINE_SUMMARY per successfully parsed restaurant, so setup is visibly confirmed without flooding the dataset. Set emitBaseline to include every item. Reuse the same monitorKey on a schedule to compare with that state.

Example input

{
  "restaurantUrls": [
    "https://famousrestaurant.toast.site/order/famous-restaurant",
    "https://restaurantconstance.toast.site/order"
  ],
  "monitorKey": "toronto-burger-competitors",
  "currencyFallback": "CAD",
  "maxPagesPerRestaurant": 1,
  "minimumPriceChangePercent": 2,
  "keywords": ["burger", "chicken", "combo"],
  "emitFullSnapshot": false
}

For Toast and other exact ordering URLs, start with one page per restaurant. Increase maxPagesPerRestaurant only when beginning from a homepage. Failed restaurants are reported in the run summary and are not charged.

Recommended schedule

Run weekly for ordinary competitor pricing, daily during promotion or seasonal-menu periods, and reuse the same monitorKey. Results are available as JSON/CSV through the default dataset and can also be delivered to an HTTPS webhook.

Recommended Store pricing

Configure two simple pay-per-event charges:

• restaurant-scan — $0.05 per successfully parsed restaurant;
• changed-restaurant — $0.05 per restaurant with one or more matched changes.

Remove the automatic apify-default-dataset-item event in the pricing setup so summary and evidence rows are not double-charged.

This prices a scan below the leading broad menu extractor while charging for the recurring state and alert layer. Failed restaurants are not charged. Offer initial Store trial credit rather than artificially billing every menu item.

Local validation

npm install
npm test
npm run smoke

The unit suite covers JSON-LD, public application-state extraction, rendered ordering rows, international prices, all change classes, severity, and filtering. In the live 10-site Toast validation, the first run parsed 7/10 menus and 998 items; the immediate second run parsed 8/10 and emitted zero raw or matched changes. A focused three-site check also parsed all three and stayed quiet on its second run.

Responsible use

Only public business-menu information is processed. Customers remain responsible for source terms, robots directives, rate limits, and applicable law.