Individuals who are blind or visually impaired struggle to identify currency notes and often rely on others for confirmation, heightening their susceptibility to fraud and diminishing their personal autonomy. This initiative suggests an independent, camera-driven currency recognition system that leverages image processing and Convolutional Neural Networks (CNN) to discern denominations and deliver voice feedback entirely without relying on IoT or cloud services. The system employs a standard camera like a smartphone or USB webcam to take a picture of the banknote. This picture goes through preprocessing stages such as converting to grayscale, filtering out noise, resizing, and performing edge detection with OpenCV. The processed image is subsequently sent through a trained CNN model that categorizes the note according to its distinct visual characteristics. Upon identification, the relevant currency value is conveyed as an audio message through an offline text-to-speech (TTS) engine, allowing the visually impaired user to comprehend the denomination without needing any visual engagement. In contrast to IoT-based systems, this approach operates fully on a local device, removing reliance on internet access and guaranteeing quicker response times. The system is lightweight, easy to use, and designed for offline access on laptops, desktops, or mobile devices. By providing precise detection and immediate feedback, this initiative presents an economical and reachable option for visually impaired people, enabling them to carry out everyday financial dealings with enhanced autonomy and assurance. It may be enhanced to incorporate multi-currency functionality and live language translation in forthcoming updates.