This paper examines the challenges visually impaired individuals face during financial transactions in the Philippines, particularly their difficulty in identifying monetary bills, which hinders their ability to manage finances independently. A solution to this issue is proposed: an image recognition application called Blind Bills. Blind Bills utilizes advanced image recognition technology with text-to-speech functionality to identify different denominations of Philippine currency, enabling visually impaired individuals to count and manage their finances accurately. The project involved capturing diverse images of currency bills, utilizing Python and machine learning algorithms for development, and conducting comprehensive testing to evaluate accuracy under various conditions. Results indicate that Blind Bills demonstrates promising accuracy rates when factors such as the number of dataset images, the distance of the camera from the bills, and the background color of the bills are properly optimized. However, exploring alternative learning models and applications for further enhancement is imperative.