Send what you know.
We'll find the car.
No mapping project, no fixed input format. Send the car as your system describes it: make and model in your own words, a VIN, a plate, catalogue IDs, a listing text or photos. We identify the vehicle and tell you which detail would make the valuation more precise.
Five ways in.
One identified vehicle.
Every input below identifies the same car. Combine them as you like: your own values always win over anything we read or look up.
Your own field names
Make, model and variant as text, in any of our markets' vocabulary.
"make": "VW", "model": "Golf", "variant": "1.5 eTSI Life DSG"A VIN or a plate
A VIN decoder, Danish plate lookup and German HSN/TSN type codes.
"vin": "WVWZZZCD…"Listing text
Up to 20,000 characters of the seller's own words.
"description": "Golf 1.5 eTSI DSG, 2021, 48.000 km, ACC, Navi"Photos
Up to eight, read for identity, equipment and condition.
"images": [ { "url": "https://…/1.jpg" }, … ]Catalogue IDs
The most exact input, from our free catalogue API.
"catalogue": { "drivetrain_id": "dv-…" }
We speak your data's language,
so you don't have to map it.
Integrations usually start with weeks of mapping one system's codes to another's. Here, the vehicle knowledge graph does that work: it knows the names, abbreviations and trade jargon sellers use for each car and feature in every market we cover.
Synonyms and local names
"VW" and "Volkswagen" are one brand; "AHK", "anhængertræk" and "gancio traino" are one towbar.
Units as you have them
Mileage in kilometres, miles or Scandinavian mil, converted to the market's unit.
Strict about typos
Unknown fields are rejected with a pointer to the field, so a misspelled key never passes silently.
Every value explained
Each value in the result says where it came from: your input, the register, text, a photo or an assumption.
Every answer says
what to send next.
Start with what you have, and the result tells you which missing input would improve it most, with a qualitative expected impact. Ask your user for that one detail, or pick it from the candidates we return, and value again.
- Hints ordered by usefulness, each with an expected impact
- Candidates to choose from when a generation or drivetrain is ambiguous
- A minimum precision, so results you can't use are refused for free
[
{
"field": "vehicle.catalogue.drivetrain_id",
"reason": "drivetrain_unresolved",
"expected_impact": "high",
"message": "Select the exact engine and
transmission from the catalogue.",
"candidates": [ … ]
},
{
"field": "vehicle.condition",
"reason": "condition_unknown",
"expected_impact": "medium"
}
]Fifteen years of cars,
mapped down to the feature.
Behind the flexible input is AutoUncle's vehicle knowledge graph: brands, models, generations, drivetrains, trims and features, linked to each other and to the words sellers use for them in every market.
See vehicle intelligence →Send a car the way
your system has it.
Free test keys, no card required. See in the console exactly how each request was identified.