Restaurant Menu Scraper

Extract structured menu data from independent and local restaurant websites. Returns clean JSON with categories, items, prices, and descriptions.

Works best with independent restaurants, local chains, and fine dining establishments that publish their own menus online.

Two Ways to Use

URL Mode (default)

Give it a restaurant URL and get back a structured menu. The scraper finds the menu on the site — whether it's an HTML page, a PDF, or listed on a third-party platform like Yelp or DoorDash — and extracts everything into clean JSON.

Search Mode

Tell it a cuisine and a city, and it discovers restaurants in that area via Google Places, then scrapes each one. Results appear in the dataset one at a time as each restaurant completes.

Input

URL Mode

Field	Type	Required	Description
mode	enum	No	Set to "url" (this is the default)
url	string	Yes	Restaurant website URL (e.g. https://franklinbbq.com)
location	string	No	For multi-location restaurants, specify which location (e.g. "San Diego")
freshness	enum	No	How fresh the data should be (default: med_cache)

Search Mode

Field	Type	Required	Description
mode	enum	Yes	Set to "search"
query	string	Yes	Cuisine or restaurant type (e.g. "sushi", "Mexican", "Italian", "BBQ")
location	string	Yes	City or area to search (e.g. "Pasadena, CA", "Austin, TX")
freshness	enum	No	How fresh the data should be (default: med_cache)

The word "restaurants" is automatically added to your search query, so just enter the cuisine type.

Freshness Tiers

Controls how recently the menu data was scraped. Cached data is faster and cheaper.

Tier	Max Age	Cost
long_cache	180 days	$0.02/request
med_cache	60 days	$0.03/request
short_cache	14 days	$0.04/request
fresh	1 day	$0.05/request

Billing

You only pay for successful, quality extractions. Failed scrapes and low-quality results are free in both URL and Search modes.

• Successful extraction with complete or mostly-complete menu → billed
• Failed extraction (site blocked, no menu found, etc.) → free
• Low-quality extraction (missing core menu sections, e.g. only drinks for a BBQ restaurant) → free

Each dataset row includes a billable field (true or false) for full transparency.

What It Extracts

Each restaurant produces a structured menu with:

• Categories — Appetizers, Entrées, Desserts, Drinks, etc.
• Menu items — Name, description, and price for each dish
• Extraction confidence — high, medium, or low so you know how complete the data is
• Completeness grade — AI-verified assessment of whether the menu is complete (see below)
• Restaurant name — The official business name as shown on the website

The scraper handles HTML menus, PDF menus, and third-party platforms (Yelp, DoorDash, etc.), automatically choosing the best source.

Completeness Grading

Every successful extraction is evaluated by AI to determine whether the menu is complete for the restaurant type. The grade appears in menu.completeness:

Grade	Meaning
complete	Full menu with all expected sections (30+ items, 4+ categories)
likely_complete	Looks complete, minor uncertainty (20+ items, 3+ categories)
partial_core_questionable	Some food items but feels thin or incomplete
partial_core_missing	Missing primary food offering (e.g., only drinks for a BBQ restaurant)
minimal	Fewer than 5 items or too vague to be useful
not_a_menu	Extracted content is not menu data

The completeness object also includes reasoning (why it got that grade), restaurant_type (detected cuisine), and missing_sections (what appears to be absent).

Important Limitations

• Fast food chains are excluded from search results. National fast food chains (McDonald's, Subway, Taco Bell, Chick-fil-A, etc.) are automatically filtered out. These large multi-location chains frequently use dynamic menu systems that don't scrape reliably, and their menus are widely available through other channels.
• Best results come from independent restaurants that publish their own menu on their website or as a PDF. Regional and fine dining chains with individual location pages also work well.
• Some restaurants will fail — not every website has a scrapeable menu. The dataset will include both successes and failures so you can see exactly what worked.

Output Examples

URL Mode

Input:

{
    "url": "https://franklinbbq.com"
}

Dataset row:

{
    "success": true,
    "url": "https://franklinbbq.com",
    "restaurantName": "Franklin Barbecue",
    "itemCount": 45,
    "confidence": "high",
    "extractionConfidence": {
        "confidence": "high",
        "item_count": 45,
        "has_prices": true,
        "has_descriptions": true,
        "price_coverage": 0.89
    },
    "menu": {
        "categories": [
            {
                "name": "Barbecue",
                "items": [
                    {
                        "name": "Brisket",
                        "price": "$28.00/lb",
                        "description": "Oak-smoked for 18 hours"
                    }
                ]
            }
        ]
    }
}

Search Mode

Input:

{
    "mode": "search",
    "query": "sushi",
    "location": "Pasadena, CA"
}

Dataset rows (one per restaurant, streamed as each completes):

{
    "success": true,
    "url": "https://sushigen.com",
    "restaurantName": "Sushi Gen",
    "searchQuery": "sushi",
    "searchLocation": "Pasadena, CA",
    "serperTitle": "Sushi Gen",
    "serperAddress": "123 Main St, Pasadena, CA 91101",
    "itemCount": 62,
    "confidence": "high",
    "extractionConfidence": {
        "confidence": "high",
        "item_count": 62,
        "has_prices": true,
        "has_descriptions": true,
        "price_coverage": 0.95
    },
    "menu": {
        "categories": [
            {
                "name": "Nigiri",
                "items": [
                    {
                        "name": "Salmon",
                        "price": "$6.50",
                        "description": "Two pieces, fresh Atlantic salmon"
                    }
                ]
            }
        ]
    }
}

Failed restaurants also appear in the dataset:

{
    "success": false,
    "url": "https://example-restaurant.com",
    "restaurantName": "Example Restaurant",
    "searchQuery": "sushi",
    "searchLocation": "Pasadena, CA",
    "error": "all_sources_failed",
    "errorMessage": "Could not extract menu from any available source"
}

Run summary (available in the OUTPUT key-value store after completion):

{
    "mode": "search",
    "query": "sushi",
    "location": "Pasadena, CA",
    "totalFound": 20,
    "excludedFastFood": 2,
    "skippedNoWebsite": 3,
    "totalScraped": 15,
    "successful": 12,
    "failed": 3
}

Use Cases

• Restaurant aggregators — Build menu databases for a city or region
• Food delivery apps — Import menus from local restaurants automatically
• Market research — Compare pricing across restaurants by cuisine and area
• Price monitoring — Track how menu prices change over time
• Location intelligence — Map restaurant offerings by geography