See what you can build

Urls: [
  "https://slickdeals.net/",
  "https://www.dealnews.com/",
  "https://www.techbargains.com/"
]
Prompt:
Search for the most popular deals from input url in the US in the last
7 days with any of included_keywords and exclude the ones with any of
excluded_keywords. Return structured JSON array for each result.

Urls: [
  "https://www.aloyoga.com/products/w9680r-airbrush-heart-throb-bra-black-white",
  "https://athleta.gap.com/",
  "https://www.aloyoga.com/products/u3031rg-accolade-crew-neck-pullover-espresso"
]
Prompt:
Go to input url and extract product details. Then search for similar products on comparison site.

Compare up to 3 products and return a ranked list with:

* product_url
* match_score (0.0-1.0)
* is_exact_match (true/false)
* reason

Return as JSON array.

Prompt:
Using input activity_name and other details, find the matching product on both input url and comparison_url. Extract:

* product name
* price
* sku or product id

Compare prices between both sites. Return as JSON.

Urls: [
  "https://www.fitnesssf.com/",
  "https://www.localsfitness.com/",
  "https://www.urbanfitnessoakland.com/"
]
Prompt:
Go to input url and extract pricing information:

* Price (as a number with currency)
* Service or product name
* Any pricing tiers or options available

Return the data as JSON.

Urls: [
  "https://beaire.com/en/aire-ancient-baths-new-york/",
  "https://thegreenwichhotel.com/wellness/shibui-spa/",
  "https://greatjonesspa.com/"
]
Prompt:
Navigate to input url for the input business name in the input city. Find pricing for service_name. Extract:

* service name (as shown on site)
* original price
* currency

Handle location selection if needed. Return as JSON.

Urls: [
  "https://joesstonecrab.com/",
  "https://www.ceviche105.com/",
  "https://www.versaillesrestaurant.com/"
]
Prompt:
Go to input url and extract all beverages from the menu:

* name
* price
* description
* variations (if any)

Also categorize as "Coke", "Pepsi", "Both", or "N/A" based on products offered. Return as JSON.

Urls: [
  "https://www.buttermilkbourbon.com/",
  "https://www.row34.com/",
  "https://midarestaurant.com/"
]
Prompt:
Visit input url and identify the online ordering system (e.g., Toast, Square, Olo). Find:

* vendor name
* evidence URL where you confirmed this

Only look for food ordering systems. Return as JSON.

Urls: [
  "https://atarayo-nishiizu.jp/",
  "http://www.osakaacademia.com/",
  "http://www.hanayamaonsen.com/"
]
Prompt:
Extract comprehensive hotel information from provided hotel URLs

Navigate to each hotel website and extract detailed information including hotel name, location, pricing, amenities, contact details, and descriptions.

## 1. Website Navigation and Verification

* Navigate to {url} provided in the input data
* Verify the website loads successfully and contains hotel-related content
* Handle redirects and ensure final destination is the correct hotel property
* Set user agent to mimic standard browser behavior
* Wait for dynamic content to load (minimum 3 seconds)

## 2. Hotel Information Extraction

* Gather amenities and services offered
* Extract pricing information when available
* Capture property descriptions and unique selling points

### A. Enhanced Pricing Information Handling

| Currency Format | Example | Standardized Output |
| -- | -- | -- |
| USD with symbol | $150/night | {"amount": 150, "currency": "USD", "period": "night"} |
| EUR with symbol | €120 per night | {"amount": 120, "currency": "EUR", "period": "night"} |
| JPY with symbol | ¥15000/泊 | {"amount": 15000, "currency": "JPY", "period": "night"} |
| Price ranges | From $96 | {"min_amount": 96, "currency": "USD", "period": "night", "type": "range"} |
| Local currency | $250,000 COP | {"amount": 250000, "currency": "COP", "period": "night"} |
| Multiple currencies | $120 USD / €110 EUR | {"primary": {"amount": 120, "currency": "USD"}, "alternate": {"amount": 110, "currency": "EUR"}} |

### B. Pricing Source Prioritization

* Primary: Hotel's official website pricing section
* Secondary: Booking platform data (Booking.com, Expedia, Hotels.com)
* Tertiary: Meta search engine results (Kayak, Trivago)
* Fallback: Rate comparison sites and travel aggregators
* Always capture currency context and convert ranges to min/max format

### C. Amenities Extraction Rules

* Scan for common hotel amenities keywords (pool, spa, wifi, parking, restaurant)
* Extract from dedicated amenities sections, icons, or service lists
* Categorize amenities by type (recreation, business, dining, transportation)
* Handle multilingual amenity descriptions
* Standardize amenity names to consistent terminology

## 4. Error Handling and Edge Cases

* Handle timeout scenarios gracefully with partial data extraction
* Manage SSL certificate errors and security warnings
* Process password-protected or members-only content areas
* Deal with JavaScript-heavy sites requiring interaction

Urls: [
  "https://www.walmart.com",
  "https://www.amazon.com",
  "https://www.target.com",
  "https://www.gamestop.com",
  "https://www.bricklink.com",
  "https://www.argos.co.uk",
  "https://www.johnlewis.com",
  "https://www.barnesandnoble.com",
  "https://www.zavvi.com",
  "https://www.kohls.com",
  "https://www.toysrus.com",
  "https://www.smythstoys.com",
  "https://www.bestbuy.com",
  "https://www.entertainmentearth.com",
  "https://www.lego.com"
]
Prompt:
Search for "{legoSetName}" Lego set on this retailer website.

TASK:
1. Look for the Lego set on this page
2. Find the specific product that matches "{legoSetName}"
3. Extract the following information:

Return JSON:
{
  "inStock": true or false,
  "price": "99.99" (number only),
  "currency": "USD",
  "shipping": "Free shipping" or "Shipping: $X.XX",
  "productUrl": "direct link to product"
}

Urls: [
  "https://carbonhealth.com/",
  "https://www.solvhealth.com/r/book-online/gqBMql/c7TeWm1J6DLIdO49wVlxN/slots/today?utm_source=direct&utm_medium=none&landing_page=https%3A%2F%2Fwww.afcurgentcare.com%2Fsunnyvale%2F",
  "https://www.elcaminohealth.org/locations/urgent-care-mountain-view",
  "https://www.sutterhealth.org/find-location/mills-peninsula-urgent-care-burlingame",
  "https://www.gohealthuc.com/ucsfhealth/locations/castro#save-my-spot"
]
Prompt:
You are looking for the EARLIEST available appointment time at this pediatric urgent care clinic.

CURRENT DATE AND TIME: Tuesday, January 14, 2025, 10:30 AM
IMPORTANT: Only return appointment times that are AFTER 10:30 AM today, or on future dates.

STEP 1 - LOCATION INPUT (if required):
If the page asks for location, address, or zip code, enter: "94086" or "Sunnyvale, CA"
Click search/find/continue to proceed.

STEP 2 - FIND BOOKING/SCHEDULING:
Look for buttons or links containing:
- "Book Online", "Schedule Now", "Reserve Spot"
- "Get in Line", "Check In Online", "Join Waitlist"
Click on it to access the scheduling system.

STEP 3 - FIND EARLIEST AVAILABLE TIME:
- Look at the calendar or time slot picker
- Find the FIRST available date (today if slots remain, otherwise next available)
- Find the EARLIEST time slot on that date

STEP 4 - RETURN RESULT:
{
  "clinic_name": "Name of the clinic",
  "earliest_date": "Today" or "01/15/2025",
  "earliest_time": "2:30 PM",
  "booking_link": "Current page URL"
}

Urls: [
  "https://www.singsaver.com.sg/credit-cards",
  "https://www.moneysmart.sg",
  "https://seedly.sg",
  "https://milelion.com",
  "https://www.suitesmile.com",
  "https://mainlymiles.com"
]
Prompt:
User wants: "{requirements}"

TASK: Extract top 3-5 credit cards from THIS PAGE ONLY that match the user's needs.

RULES:
1. Stay on this page - do NOT click into individual card pages
2. Extract info visible on the listing/comparison page
3. Be quick - just get the key details shown

Return JSON:
{
  "cards": [
    {
      "name": "Card Name",
      "issuer": "Bank",
      "annualFee": "S$XXX or No fee",
      "rewards": "Brief rewards summary",
      "signUpBonus": "Bonus or null",
      "apr": "XX% or null",
      "highlights": ["key benefit 1", "key benefit 2"]
    }
  ]
}