Car Listing Condition Grader

What does Car Listing Condition Grader do?

Car Listing Condition Grader turns a used-car listing URL into a graded vehicle condition report. Paste an AutoScout24 listing or search URL and the Actor pulls the listing's own photos, inspects them with AI vision, and returns a condition grade (A–E), a 0–100 condition score, and a panel-by-panel list of visible damage — dents, scratches, rust, cracks, missing parts — alongside interior wear, tyre condition, and photo-quality warnings.

Most vehicle damage tools require you to already have the photos. This one starts from the live listing: listing URL in, graded vehicle out.

Important — this is an AI estimate from photographs, not a mechanical inspection. The Actor can only judge what is visible in the seller's own photos. It cannot detect engine, gearbox, electrical, structural or underbody problems, accident history, odometer fraud, or anything a seller chose not to photograph. Output is non-deterministic: the same listing can score slightly differently across runs. Use it to triage and rank listings at scale, then inspect the shortlist properly. Never treat a grade as a substitute for a professional inspection or a purchase decision on its own.

Why grade car listings automatically?

• Auction and wholesale buying — Triage hundreds of lots before bidding and filter out the wrecks instead of opening every listing by hand.
• Dealer sourcing — Rank trade-in and remarketing inventory by visible condition before you send a buyer to look.
• Listing quality control — Spot stock photos, missing interior shots and watermarked galleries across your own inventory.
• Price-versus-condition analysis — Join condition grades to price and mileage to find underpriced clean cars and overpriced rough ones.
• Damage datasets — Build structured, labelled damage records for analytics or model training.

How to use Car Listing Condition Grader

1. Open the Actor in Apify Console.
2. Add one or more AutoScout24 search or listing URLs under Start URLs. Search pages are expanded into individual cars automatically.
3. Set Max vehicles per start URL and Max photos per vehicle.
4. Run the Actor and download results from the Dataset tab.

No listing page? Put photo URLs straight into Direct photo URLs instead — see below.

Input

Field	Type	Default	Description
startUrls	Array	—	AutoScout24 detail URLs (/angebote/…) or search URLs (/lst/…). Country domains .de, .at, .it, .nl, .fr, .be, .es, .lu are supported.
maxItems	Integer	5	Maximum graded vehicles per start URL.
maxPhotosPerVehicle	Integer	8	Maximum photos analyzed per car (1–20).
imageUrls	Array of strings	—	Optional. Direct photo URLs for a single vehicle — see Direct photo mode.
proxyConfiguration	Object	no proxy	Optional proxy settings.

{
  "startUrls": [
    { "url": "https://www.autoscout24.de/lst/bmw/3er?atype=C&cy=D&sort=standard" }
  ],
  "maxItems": 5,
  "maxPhotosPerVehicle": 8,
  "proxyConfiguration": { "useApifyProxy": false }
}

Direct photo mode

Grade a car whose listing lives on a portal this Actor does not support, or whose photos you already store yourself:

{
  "imageUrls": [
    "https://example.com/car/front.jpg",
    "https://example.com/car/rear.jpg",
    "https://example.com/car/interior.jpg"
  ],
  "maxPhotosPerVehicle": 8
}

All URLs in imageUrls are treated as photos of the same vehicle and produce one graded row, with listingUrl and the vehicle specification fields set to null.

Output

One row per vehicle. Download as JSON, CSV, Excel, or HTML.

{
  "listingId": "58d557c5-253d-44e4-921b-8e75f33246f0",
  "listingUrl": "https://www.autoscout24.de/angebote/volkswagen-sonstige-golf-v-us-modell-...",
  "title": "Volkswagen Sonstige Golf V US Modell Verkaufe gegen Gebot",
  "make": "Volkswagen",
  "model": "Sonstige",
  "year": 2007,
  "price": 1,
  "currency": "EUR",
  "mileage": 225000,
  "imageUrl": "https://prod.pictures.autoscout24.net/listing-images/58d557c5-..._66543e8b-....jpg/1280x960.webp",
  "imageUrls": ["https://prod.pictures.autoscout24.net/listing-images/..."],
  "photosAnalyzed": 6,
  "conditionGrade": "D",
  "conditionScore": 50,
  "damages": [
    { "panel": "sill-left", "type": "rust", "severity": "moderate", "confidence": 1 },
    { "panel": "sill-right", "type": "rust", "severity": "moderate", "confidence": 1 }
  ],
  "interiorWear": null,
  "tyreCondition": "good",
  "photoFlags": ["interior-missing", "dirty-vehicle"],
  "summary": "The vehicle exhibits significant, visible corrosion along both the left and right side sills; interior photos are missing.",
  "scrapedAt": "2026-08-01T07:41:12.903Z"
}

Field	Description
listingId	Listing identifier on the source portal — null in direct photo mode
listingUrl	Canonical listing URL — null in direct photo mode
title	Vehicle title built from make, model and trim
make / model	Manufacturer and model group
year	Year of first registration
price / currency	Asking price as a number, and its ISO currency code
mileage	Odometer reading in kilometres
imageUrl	Primary listing photo (absolute URL) — null when the listing has no photos
imageUrls	Full photo gallery in listing order
photosAnalyzed	How many photos were sent to AI vision for this car
conditionGrade	A–E, derived from conditionScore so the two never disagree
conditionScore	Visible condition from 0 to 100
damages	Visible defects: panel, type, severity (minor/moderate/severe), confidence (0–1)
interiorWear	excellent/good/fair/poor, or null when no interior photo is clear enough to judge
tyreCondition	excellent/good/fair/poor, or null when the tyres are not clearly visible
photoFlags	Reasons the grade may be limited (see below)
summary	One-sentence assessment of the visible condition
scrapedAt	ISO timestamp of the grading

Grade scale

Grade	Score	Meaning
A	90–100	Showroom condition, no visible defects
B	75–89	Good — light wear only, or photos too limited to certify perfection
C	55–74	Average — several visible cosmetic defects
D	35–54	Poor — significant damage, corrosion or a cracked panel
E	0–34	Wrecked, crash-damaged or a repair project

Listings with no usable photos are still saved, with conditionGrade and conditionScore set to null and photoFlags containing no-photos.

Photo flags

photoFlags tells you how much to trust the grade. A clean car photographed from three distant angles is not the same evidence as one photographed close up from twelve.

stock-photo · low-resolution · plate-obscured · interior-missing · exterior-incomplete · wheels-not-visible · heavy-editing · poor-lighting · dirty-vehicle · obstructed-view · watermarked · non-vehicle-photo · detail-too-limited · no-photos

Pricing

Pay-per-event — see the Store pricing tab for current rates.

Event	When charged
listing-scraped	Once per listing page read, including listings that turn out to have no usable photos
photo-analyzed	Once per photo sent to AI vision, capped by maxPhotosPerVehicle
vehicle-graded	Once per graded vehicle written to the dataset

Budgeting tip: photo-analyzed dominates the bill — a run at maxPhotosPerVehicle: 8 costs about four times as much per car as one at 2. But photo count drives result quality too. At 2 photos a car is usually only graded on one or two exterior angles, so damages comes back empty and interiorWear and tyreCondition come back null on most listings — enough to spot an obvious wreck, not enough to rank clean cars against each other. Use 6–8 photos when you need kerb rash, panel defects and interior wear. All photos of one car are sent in a single AI call regardless of the setting.

Notes and limitations

• Photos only. Mechanical condition, service history, accident records and hidden rust are invisible to this Actor.
• The seller controls the evidence. Sellers photograph the good side. Missing angles are reported through photoFlags, not guessed at.
• Non-deterministic output. Grades can shift by a few points between runs on the same listing. Rank and filter on the grade; do not treat the score as an exact measurement.
• Public data only. The Actor reads publicly visible listing pages. It never logs in to any account.
• Listings expire. Used-car listings are removed quickly; refresh your start URLs for repeat runs.

Image Credit

Image credit: autoscout24.de