ANPR · ALPR · LPR — Real-time plate intelligence

License plates,
recognized in milliseconds.

Wyom turns any camera into a precision Automatic Number Plate Recognition system — 99.7% accuracy, 140+ countries, edge or cloud. Built for parking, tolling, security and smart cities.

Try the live playground Read the API docs
99.7%
accuracy
800
median latency
0
countries
2.4+
plates read

Powering plate intelligence for

  • MetroPark
  • TollBridge Group
  • CityGrid
  • Sentia Security
  • LogiFleet
  • Aurora Municipal

Product

Everything an ANPR deployment needs

From a single camera at a gate to 10,000 lanes across a continent — one platform, edge to cloud.

Real-time recognition

Sub-40ms inference per frame at the edge. Plates are read, matched and actioned before the bumper crosses the line.

140+ country formats

EU, US, UK, GCC, APAC and more — reflective, embossed, two-line and motorcycle plates handled natively.

Edge AI cameras

ONVIF/RTSP compatible. Run Wyom on your existing cameras or deploy our IP67-rated edge units with on-device OCR.

Hotlists & alerts

Match against watchlists, VIP lists and permit databases in real time with webhook, SMS and dashboard alerts.

Vehicle analytics

Make, model, color, direction, dwell time and traffic heatmaps — GDPR-compliant with automatic face & plate blurring.

Developer-first API

REST and WebSocket APIs, SDKs for Python, Node and Go, and a sandbox playground. Ship an integration in an afternoon.

How it works

From pixels to plate in three steps

  1. 01

    Capture

    Any camera streams frames to Wyom over RTSP, or our edge units process video on-device with zero bandwidth cost.

  2. 02

    Recognize

    Deep-learning detection finds the plate region; a dedicated OCR head reads characters with per-character confidence scoring.

  3. 03

    Act

    Results match against your lists and trigger gates, barriers, alerts, billing events or analytics pipelines — instantly.

Simulation

Watch Wyom read a plate

Type a plate — or roll the dice — and see a simulated live recognition event end-to-end.

CAM-04 · Gate North · 1080p @ 30fps

Letters and digits only — simulates an in-the-wild capture.

Plate—
Confidence
—
Country—
Timestamp—
Decision—

Developers

An API your engineers will love

One REST call. Plate text, bounding box, per-character confidence and vehicle attributes back in ~40ms.

# Recognize a plate from an image
curl -X POST https://parking.wyom.in/ocr \
  -H "Authorization: Bearer wy_live_••••••••" \
  -F "image=@capture.jpg"
from wyom import WyomClient

client = WyomClient(api_key="wy_live_...")
result = client.plates.recognize("capture.jpg")

print(result.plate)        # "UAE Dubai A 42352"
print(result.confidence) # 0.994
import { Wyom } from "@wyom/sdk";

const wyom = new Wyom({ apiKey: "wy_live_..." });
const { plate, confidence } = await wyom.plates.recognize("capture.jpg");
# Stream frames, receive events
wscat -c wss://stream.wyom.in/v1/live \
  -H "Authorization: Bearer wy_live_..."

# → {"plate":"UAE Dubai A 42352,"confidence":0.994,
#    "camera":"cam-04","direction":"in"}
200 OK · application/json
{
  "plate": "UAE Dubai A 42352",
  "confidence": 0.994,
  "details": { "region": "Dubai", "code": "A", "country": "UAE" },
  "box": { "x": 512, "y": 388, "w": 214, "h": 54 },
  "vehicle": { "type": "sedan", "color": "silver", "model": "Nissan Patrol" },
  "latency_ms": 38
}
  • REST + WebSocket — batch images or stream live video
  • SDKs — Python, Node.js, Go, .NET
  • Webhooks — signed events for matches & alerts
  • 99.99% SLA — multi-region failover on paid tiers

Pricing

Simple, scalable pricing

Start free. Scale to billions of recognitions. No per-camera lock-in.

Developer

$0/mo

  • 1,000 recognitions / month
  • REST API + playground
  • Community support
  • 1 camera connection
Start free
Most popular

Pro

$49/mo

  • 100,000 recognitions / month
  • Edge + cloud deployment
  • Hotlists, webhooks & alerts
  • 25 camera connections
  • Priority email support
Start 14-day trial

Enterprise

Custom

  • Unlimited recognitions
  • On-prem & air-gapped installs
  • SSO, audit logs, DPA
  • Dedicated success engineer
  • 99.99% uptime SLA
Talk to sales

About Wyom

We believe moving vehicles shouldn't move data slowly

Founded in 2026 by computer-vision researchers, Wyom started as a parking-gate hackathon project. Today our models read 2.4 billion plates a year across 140 countries — for tolling operators, city governments, logistics fleets and security teams who can't afford to guess.

We train on consented, privacy-safe datasets, blur faces by default, and ship models small enough to run on a $60 edge device. Recognition you can trust, infrastructure you can own.

  • Privacy-first — GDPR & CCPA compliant, on-prem options
  • Radically reliable — 99.99% uptime across 3 regions
  • Human support — engineers, not ticket queues
0Founded
0Countries deployed
0BPlates / year
0%Uptime SLA

FAQ

Questions, answered

What is ANPR / ALPR / LPR?

Automatic Number Plate Recognition (also ALPR or LPR) uses computer vision and OCR to detect and read vehicle license plates from images or video in real time — powering parking, tolling, security and traffic analytics.

How accurate is Wyom?

99.7% character-level accuracy across 140+ countries, including reflective, damaged, angled and low-light plates. Every read carries per-character confidence so you can set your own thresholds.

Does it work with my existing cameras?

Yes — Wyom connects to any ONVIF/RTSP camera. We also offer IP67-rated edge AI cameras with on-device processing when you need zero-bandwidth operation.

Is there a free trial?

The Developer plan is free forever with 1,000 recognitions per month — no credit card required. Paid plans include a 14-day trial.

How do you handle privacy & GDPR?

Face and plate blurring is on by default, retention windows are configurable per deployment, and Enterprise supports fully air-gapped on-prem installs with no external calls.

Contact

Let's read some plates together

Tell us about your deployment — parking, tolling, security, smart city — and we'll get back within one business day.