AutoScout24 Scraper

Scrape car listings from AutoScout24: price, mileage, first registration, power, fuel, transmission, damage flag, dealer and location — from any search you can build on the site.

Built around one rule the alternatives break:

A result that does not match your filter is worse than no result at all.

The problem this Actor solves

Ask AutoScout24 for a make it does not know — a typo, a renamed brand, a model that moved — and it answers HTTP 200 with a full page of cars. Not an error. Not an empty list. The entire marketplace, 800 000+ listings, sorted as if you had asked for them.

A scraper that trusts the page will hand you thousands of rows that look perfectly fine and have nothing to do with your request, and bill you for every one.

This Actor checks that AutoScout24 actually applied your filter before it writes a single row. If it didn't, the run stops and tells you which part of your input was not recognised.

Situation	What happens	You pay
Make or model not recognised	Run succeeds, dataset empty, the unrecognised value is named	Nothing
Search matches zero cars	Run succeeds, reason explained	Nothing
AutoScout24 blocked us or changed its page structure	Run fails loudly	Nothing
Listings found	Run succeeds with data	Per result

Input

Either paste search URLs:

{ "startUrls": ["https://www.autoscout24.com/lst/bmw/x1?cy=D&priceto=15000"], "maxItems": 200 }

or describe the search:

{
  "make": "bmw",
  "model": "x1",
  "country": "D",
  "priceTo": 15000,
  "yearFrom": 2015,
  "mileageTo": 150000,
  "sort": "price",
  "maxItems": 200
}

Output

One row per listing:

{
  "listingId": "e8854adb-97eb-466a-bfcb-f34fb6e7a1a2",
  "url": "https://www.autoscout24.com/offers/bmw-x1-18-d-sdrive-diesel-...",
  "make": "BMW",
  "model": "X1",
  "variant": "18 d sDrive",
  "priceEur": 2900,
  "mileageKm": 242000,
  "firstRegistrationYear": 2010,
  "firstRegistrationMonth": 11,
  "powerKw": 105,
  "powerHp": 143,
  "fuel": "Diesel",
  "transmission": "Manual",
  "engineCcm": 1995,
  "isDamaged": true,
  "sellerType": "Dealer",
  "sellerName": "Autohandel Younes K.",
  "country": "DE",
  "zip": "54311",
  "city": "Trierweiler",
  "imageCount": 13,
  "scrapedAt": "2026-08-19T17:00:00.000Z"
}

Numbers come out as numbers. 242,000 km becomes 242000, 105 kW (143 hp) becomes two fields.

A RUN_REPORT record in the key-value store records every search, what came back, and — if anything went wrong — which side it was on.

Known limit, stated up front

AutoScout24 serves at most 200 pages per search, roughly 4000 listings, no matter how many results it claims. When you ask for more, this Actor says so in the log and in RUN_REPORT instead of quietly stopping. Split the search by price band or registration year to go past it.

Use cases

• Price benchmarking for a specific make, model and year
• Dealer inventory monitoring
• Residual value and depreciation analysis
• Sourcing damaged or high-mileage cars
• Feeding a used-car pricing model

Tests

npm install
npm test

22 tests, no network needed — including the one that matters: an unrecognised make must never turn into a full marketplace dump.

Need a scraper for something else?

I build custom Apify Actors and browser automation the same way this one is built: a verified result or an explicit failure, never an empty dataset sold as a success.

• My other Actors: https://apify.com/marekhartmann
• Code and test suites: https://github.com/marekhartmann-creator

Tell me the site and what you need out of it.