One API call turns a vehicle photo into structured data — plate number, car brand, model, and color. Get a key, top up a wallet, and ship. No sales calls, no seat licenses, no minimum contracts.
150 free API calls per month. No credit card required.
Generic ANPR models are trained on clean, front-on plate crops. Then they meet a stacked-character truck plate behind a tinted cover at dusk — and guess.
Why generic ANPR fails
How we built for it
Three steps from photo to structured data — no ML expertise required on your side.
01
POST a photo to the detect endpoint — from a camera snapshot, a phone, or frames pulled off your existing CCTV.
02
A pipeline of specialized models runs detection, classification, and OCR — plates localized and read, the vehicle identified.
03
Plate text, confidence scores, car brand, model, body type, and color — back in under 200 milliseconds.
Six markets, each with its own formats and edge cases — covered in depth. Every country page details plate formats, use cases, and integration.
Everything the system extracts from a single vehicle photograph — delivered as clean JSON.
Plates are localized, cropped, and read in a single pass — day or night, at an angle, in motion.
Beyond the plate — the same request classifies the vehicle itself for richer matching and audit trails.
A REST API you can call from anything — or the full system deployed on your own infrastructure.
Under the Hood
Every image passes through a chain of specialized models. Vehicles are detected first, plates are localized and read via OCR, and the vehicle itself is classified by brand, model, and color — all within a single request.
< 200ms
Inference per image
97%+
Plate read accuracy
2,003
Car models recognized
Vehicle Detection
DetectionLocalize all vehicles in frame using real-time object detection.
License Plate Localization
DetectionCrop and align the plate region from each detected vehicle.
Plate OCR
RecognitionRead characters from the plate image — including stacked and vanity layouts.
Car Brand & Model Classification
ClassificationIdentify make, model, body type, and color from the vehicle photograph.
No subscriptions, no seat licenses. Top up a wallet and pay per recognition — the rate drops as your volume grows.
Every account starts with 150 free API calls per month. No credit card required.
See full pricing details →$0.02per request
Up to 10,000 plates / month
$100 minimum top-up
$0.015per request
10,000 – 50,000 plates / month
$150 minimum top-up
$0.01per request
50,000 – 200,000 plates / month
$500 minimum top-up
from $0.007per request
200,000+ plates / month
Custom terms
Anywhere a vehicle passes a camera and a decision follows — gate access, billing, enforcement, or audit. The same API powers all of them.
One curl command to your first plate read. Full reference docs and camera integration guides when you go deeper.
# Read a plate with one request
curl -X POST https://api.example.com/v1/detect \
-H "Authorization: Bearer YOUR_API_KEY" \
-F "[email protected]"
# Response
{
"success": true,
"plates_found": 1,
"plates": [
{
"text": "7TKR952",
"confidence": 0.98,
"position": { "x1": 412, "y1": 690, "x2": 578, "y2": 734 }
}
],
"car": {
"brand": "Toyota",
"model": "Camry",
"body_type": "sedan",
"color": { "name": "white", "hex": "#F2F3F5" }
},
"response_time_ms": 142
}Everything you need to know before your first API call.
Yes. We offer fully offline, on-premise deployment of the recognition models — Docker containers, air-gapped environments, and dedicated GPU hardware are all supported. It is available on Enterprise terms; contact us and we will scope your setup.
Still have questions? Contact us
Get an API key and run your first recognition in minutes — or send us sample images from your cameras and we will share the results.
150 free API calls per month. No credit card required.
Contact
Cameras, volumes, countries, timeline — send us a note and we will reply within one business day.