ApolloDuck Scraper & Market Intelligence

Search ApolloDuck properly instead of clicking through categories by hand.

This actor scrapes public ApolloDuck listings and returns them as structured data. If you enable AI, it can also turn a plain-English request into filters, rank the results by relevance, and write a report for a buyer, broker, or seller.

It was built for ApolloDuck specifically, which matters because ApolloDuck is not just another yacht marketplace. It has sailing yachts and motorboats, but also narrowboats, barges, houseboats, inland craft, fishing boats, inflatables, and other categories that generic marine scrapers usually handle badly.

Why use this actor

• Search ApolloDuck with either structured filters or a plain-English query
• Pull listing data into a clean dataset instead of working from inconsistent page HTML
• Rank results by how well they match your criteria
• Scrape detail pages for more useful specs, descriptions, seller information, and photos
• Generate a report when you want help comparing listings or thinking about the market

What it is good for

• Buyers trying to narrow down live listings quickly
• Brokers who want a snapshot of inventory and pricing
• Sellers who want rough comparable asking-price context
• Researchers building ApolloDuck datasets without writing a custom scraper

Typical examples:

• "Used narrowboats in the UK under £80k"
• "Family sailing catamarans in Greece or Croatia under €250k"
• "Barges in France with detail pages and pricing data"
• "What does the current narrowboat market look like?"

How it works

There are two ways to use it.

1. Structured filters

Use this when you already know the filters you want.

{
  "boatType": ["narrowboat", "barge"],
  "country": ["United Kingdom", "France"],
  "priceMax": 150000,
  "yearMin": 1995,
  "scrapeDetails": true,
  "maxResults": 50
}

2. Natural-language search

Use this when you want the actor to interpret the request for you.

{
  "query": "57ft narrowboat in the UK under £90k, good liveaboard layout",
  "mode": "buyer",
  "enableAI": true,
  "maxResults": 30
}

Seller-style example:

{
  "query": "What is my 2012 Dutch barge in France worth?",
  "mode": "seller",
  "enableAI": true,
  "scrapeDetails": true,
  "maxResults": 40
}

Main input fields

Search

• query: plain-English request
• boatType: array of boat types
• country: array of country names
• brand: optional manufacturer / builder filter
• priceMin, priceMax: price range
• yearMin, yearMax: build year range
• lengthMin, lengthMax: length range in metres
• condition: new, used, or any
• sortBy: result ordering
• maxResults: hard cap for the scrape

Intelligence

• enableAI: turns on query parsing and report generation
• mode: buyer, seller, or broker
• aiApiKey: optional user-supplied API key
• aiBaseUrl: optional OpenAI-compatible endpoint
• aiModel: optional model override

Scraping depth

• scrapeDetails: visits each detail page for richer specs and images

For broad inventory scans, leave scrapeDetails off.

For smaller, more important searches where equipment and detail matter, turn it on.

Supported boat types

The schema currently supports:

sailboat, sailing-yacht, motorboat, motor-yacht, catamaran, houseboat, fishing-boat, RIB, jet-ski, superyacht, trawler, flybridge, cabin-cruiser, center-console, bowrider, trimaran, classic-sailing-yacht, classic-power-boat, pontoon-boat, dinghy, narrowboat, barge, sport-boat, inflatable-boat

These are mapped internally onto ApolloDuck's actual categories and subcategories.

What comes back

Each run pushes one dataset item containing:

• report
• query
• mode
• totalListings
• priceStats
• listings
• processingSeconds

Each listing can include:

• title and URL
• original price and normalized EUR price
• boat type, brand, model
• year, length, beam
• location and country
• seller and seller type
• material and engine
• images
• relevance score
• matched and missing features
• detail-page data when scrapeDetails is enabled

AI behavior and billing

enableAI controls the intelligence layer.

When AI is on, the actor can:

• parse a natural-language query into structured filters
• generate an AI report after scraping results

Current billing behavior is:

• listing-scraped is charged once per listing returned
• report-generated is charged only when a report is successfully generated using the actor's built-in AI key
• if the user supplies their own API key, the report is not charged
• if no report is generated, nothing is charged for the report
• if no listings are found, no report is generated

If you want raw scraping only, set enableAI to false.

Practical usage advice

For buyers

Keep maxResults relatively low and turn scrapeDetails on when equipment or layout matters.

For brokers

Pull a wider result set first, then run more focused searches if needed.

For sellers

Do not over-constrain the price range. Comparable asking prices are more useful when the actor can see the wider market.

Notes and limitations

• ApolloDuck is an asking-price marketplace, not a sold-prices database
• Price normalization uses static exchange rates in code, not live FX
• Detail scraping is slower because each listing page is opened individually
• Some listings are sparse, so the output quality depends on what ApolloDuck actually shows
• Sold or sale-agreed listings are excluded when that status is detectable

In plain terms

If you want ApolloDuck data in a usable format, this actor will do that.

If you also want help interpreting the results, turn AI on and use the report mode that fits the job.