MareArts ANPR SDK
MareArts ANPR — SDK & Server
Automatic Number Plate Recognition — detection + OCR for 80+ countries, GPU acceleration, ARM/Windows support, and cloud Vehicle Intelligence (make, model, color, type, face, plate nation, server-side OCR).
🐍 Python SDK
Detect and read license plates in 5 lines of Python.
Installation
# CPU (all platforms: Linux, macOS, Windows, ARM)
pip install marearts-anpr
# GPU — use the built-in setup command:
ma-anpr gpu-setup cuda # NVIDIA CUDA
ma-anpr gpu-setup directml # Windows AMD/Intel/NVIDIARequirements: Python 3.10 — 3.14
Quick Start
from marearts_anpr import (
ma_anpr_detector_v16, ma_anpr_ocr_v16, marearts_anpr_from_image_file
)
detector = ma_anpr_detector_v16("640p_fp32", user_name, serial_key, signature)
ocr = ma_anpr_ocr_v16("fp32", "univ", user_name, serial_key, signature)
result = marearts_anpr_from_image_file(detector, ocr, "car.jpg")
print(result)
# {'results': [{'ocr': 'ABC1234', 'ocr_conf': 99, 'ltrb': [...], ...}], ...}V16 Models
V16 ships one unified model for detection and one for OCR — no model-size selection needed.
| Component | Model | Precision | Notes |
|---|---|---|---|
| Detector | 320p_fp32 |
fp32 / int8 | Fast — ~2× speed of 640p |
| Detector | 640p_fp32 |
fp32 / int8 | High detection — best for distant/small plates |
| OCR | fp32 |
fp32 / int8 | Single model, all regions, dynamic switching |
Dynamic Region Switching
Switch OCR region instantly without reloading models — supports 80+ countries:
ocr.set_region('eu') # Europe+ (37 countries)
ocr.set_region('kr') # Korea
ocr.set_region('na') # North America
ocr.set_region('cn') # China
ocr.set_region('jp') # Japan
ocr.set_region('univ') # Universal (default)Pass a 2-letter country code (e.g. de, au, th) for best accuracy, or use a group code. Unknown codes fall back to univ.
Input Formats
import cv2
from PIL import Image
from marearts_anpr import marearts_anpr_from_cv2, marearts_anpr_from_pil
result = marearts_anpr_from_image_file(detector, ocr, "car.jpg")
result = marearts_anpr_from_pil(detector, ocr, Image.open("car.jpg"))
result = marearts_anpr_from_cv2(detector, ocr, cv2.imread("car.jpg"))CLI
ma-anpr car.jpg # detect + OCR
ma-anpr car.jpg --region kr # with specific region
ma-anpr car.jpg --backend cuda # GPU acceleration
ma-anpr *.jpg --json results.json # batch processing
ma-anpr config # configure credentials
ma-anpr validate # validate license📖 Full SDK Documentation & Examples →
🚗 Vehicle Intelligence (Cloud AI)
Go beyond plate text — identify the vehicle itself with a single API call.
| Feature | Example Output |
|---|---|
| Make | Toyota, BMW, Hyundai, … |
| Model | Camry, 3 Series, Tucson, … |
| Color | White, Black, Silver, … |
| Type | Sedan, SUV, Truck, Van, … |
| Face | front, rear |
| Plate Nation | KR, DE, US, … |
| Server-side OCR | Cross-check plate text via cloud |
from marearts_anpr import ma_anpr_mmc
mmc = ma_anpr_mmc(user_name, serial_key, signature)
# Pass mmc to any detection call
result = marearts_anpr_from_image_file(detector, ocr, "car.jpg", mmc)
# Each result now has 7 mmc fields:
for r in result["results"]:
print(r["ocr"], r["mmc_make"], r["mmc_model"], r["mmc_color"])
# ABC1234 Toyota Camry WhiteAlso available via the REST API server — just call the /api/anpr/mmc endpoint.
🌐 REST API Server
Production-ready server with web dashboard — zero code deployment. Included in pip install marearts-anpr.
Quick Start
ma-anpr config # one-time credential setup
ma-anpr server start # http://localhost:8000
ma-anpr server start --daemon # run in background# Detect plates via REST API
curl -X POST http://localhost:8000/api/anpr -F "image=@car.jpg"
# With Vehicle Intelligence
curl -X POST http://localhost:8000/api/anpr/mmc -F "image=@car.jpg"Web Dashboard
20+ REST API Endpoints
| Category | Endpoints |
|---|---|
| Detection |
POST /api/anpr · POST /api/anpr/mmc · batch processing |
| History | List, search, detail, delete detections · CSV/JSON export |
| Watchlist | Add, list, delete watched plates · alert counts |
| Config | Region, threads, models, MMC status |
| Monitor | Health, stats, event stream (SSE), logs |
Server CLI
ma-anpr server start # foreground
ma-anpr server start --daemon # background
ma-anpr server start --port 9000 # custom port
ma-anpr server stop # stop server
ma-anpr server stop --all # stop all instances
ma-anpr server list # list running instances
ma-anpr server status # server info
ma-anpr server logs --follow # tail logs
# Detection via CLI (server must be running)
ma-anpr server detect car.jpg # local ANPR only
ma-anpr server detect car.jpg --mmc # ANPR + Vehicle Intelligence
ma-anpr server health # health check
ma-anpr server mmc # MMC status and quota🚦 Road Object Detection
Detect persons, vehicles, and 2-wheelers in real-time — included in your ANPR license at no extra cost.
from marearts_road_objects import ma_road_object_detector
rod = ma_road_object_detector("640p_fp32", user_name, serial_key, signature)
result = rod.detect("street.jpg")
# [{'class': 'car', 'conf': 0.97, 'ltrb': [...]}, {'class': 'person', ...}]Use alongside ANPR for complete traffic scene understanding — plate reading + vehicle/pedestrian awareness in one pipeline.
📖 Road Objects Documentation →
🌍 Supported Regions
80+ countries across 12 regional groups with per-country character sets for maximum accuracy.
| Region | Countries | Code |
|---|---|---|
| 🇪🇺 Europe+ | 37 countries (EU + Balkans + Indonesia) | eu |
| 🇷🇺 Ex-USSR | 15 countries (Russia, Ukraine, Kazakhstan, …) | ru |
| 🌏 Asia | 17 countries (Japan, Thailand, Vietnam, …) | asia |
| 🇺🇸 North America | USA, Canada, Mexico | na |
| 🇧🇷 South America | Brazil, Argentina | sa |
| 🇿🇦 Africa | South Africa, Nigeria | af |
| 🇦🇺 Oceania | Australia, New Zealand | oc |
| 🇬🇧 UK | England, Scotland, Wales, N. Ireland | uk |
| 🇨🇳 China | All provinces | cn |
| 🇰🇷 Korea | All plate types | kr |
| 🇯🇵 Japan | All prefectures | jp |
| 🌍 Universal | All of the above | univ |
Pass a 2-letter ISO country code (e.g. de, au, th), a full name (germany, australia), or a group code. Unknown codes fall back to univ.
📖 Full Region & Country Reference →
📱 Mobile App
On-device ANPR for iOS and Android — parking, security, fleet management.
- On-device AI — ~100-160 ms, 100% offline capable
- Team Work — leader/member collaboration with shared data
- Webhooks — Discord, Slack, or any custom endpoint
- Cloud Sync — sync across devices via marearts.com
- Vehicle Intelligence — make, model, color, type, face, plate nation, server-side OCR
- Map View, CSV Export, Whitelist/Blacklist, Stats
No additional license required — the app works as your ANPR license.
MareArts Ecosystem
- marearts-anpr — Automatic Number Plate Recognition (GitHub)
- marearts-road-objects — Road object detection (GitHub)
- marearts-xcolor — Color extraction and similarity (GitHub)
- marearts-mast — Real-time panoramic stitching (GitHub)
Support & Resources
| Resource | Link |
|---|---|
| 📧 Contact | hello@marearts.com |
| 🏠 Homepage | https://marearts.com |
| 💳 License | ANPR Solution |
| 🎮 Live Demo | http://live.marearts.com |
| 📺 Videos | YouTube Playlist |
| 📱 App Store | MareArts ANPR App |
| 🤖 Google Play | MareArts ANPR App |
License
This software requires a valid license key. Visit MareArts ANPR Solution for licensing options.
© 2026 MareArts. All rights reserved.