A precise, offline-capable ALPR SDK with fully on-device recognition. No image or plate data is sent to the cloud, with no per-scan fees.
A consistent API across iOS, Android, and Flutter so your team integrates once and ships everywhere.
Native XCFramework using Apple Vision and AVFoundation. SwiftUI and UIKit compatible. Ships as a binary framework — no source required.
Available nowCameraX + ML Kit powered AAR library. Jetpack Compose and View-based apps both supported. Minimum Android 8.0 (API 26).
Available nowA first-class Flutter plugin wrapping the native iOS and Android SDKs. Drop-in widget, event streams, and full configuration API.
Available nowAll inference runs on-device using the platform's native ML stack. Only anonymous aggregate licence and usage counters synchronize periodically when connectivity is available.
Region-aware validation with per-country regex patterns. Supports standard and vanity formats for US, UK, EU, India, Australia, Middle East, and more.
Frame-by-frame scanning at full camera frame rate. Crops the pixel buffer to your defined region of interest before OCR — no wasted computation.
API key + secret key validation against your portal. Enforce app-level restrictions, usage tiers, and expiry dates without resubmitting to the App Store.
Activate with your API key, set a delegate, call startScanning(). The SDK handles camera permissions, session management, Vision requests, cropping, validation, and deduplication. Drop in the preview view and you're done.
// 1. Activate your licence PlateVisionSDK.shared.activate(apiKey: "pv_live_xxx", secretKey: "sk_xxx") { result in if case .success(let licence) = result, licence.isValid { // 2. Configure (optional) var config = PVConfiguration.default config.region = .us config.minimumConfidence = 0.70 PlateVisionSDK.shared.configure(config) // 3. Set delegate and start PlateVisionSDK.shared.delegate = self PlateVisionSDK.shared.startScanning() } } // 4. Receive results func plateVision(_ sdk: PlateVisionSDK, didDetect result: PVDetectionResult) { print(result.plateText, result.confidencePercent) }