I need a turnkey ANPR solution that watches the cameras at both my entrance and exit lanes, recognises every plate—regular vehicles as well as emergency units—and fires off an instant, time-stamped notification to me on Telegram and WhatsApp simultaneously. Feed: one IP cameras already in place (RTSP). Expectation: the system detects the plate, confirms its characters with high confidence, tags the direction (IN or OUT), and sends a concise message such as “IN • MH12AB1234 • 2024-06-04 14:21:08”. I’m open to any solid stack—OpenCV, YOLO, EasyOCR, OpenALPR, PlateRecognizer API, or a custom-trained model—as long as it runs reliably on a modest Linux box (Ubuntu 22.04) and can be containerised for easy redeployment. Deliverables • Source code and model files • A small setup script or Dockerfile • Telegram and WhatsApp bot integration keys abstracted to a config file or .env • README explaining camera configuration and how to add more chat recipients Acceptance will be based on live tests from both camera feeds showing >95 % accuracy in daylight and usable detection after dusk with IR.