Approximately 262 million people worldwide suffer from asthma, highlighting the critical need for early detection of asthma attacks and the utmost importance of correct inhaler usage. To address this issue, BreatheSmarter was created, which is a highly economical and efficient smart attachment for any standard inhaler that helps patients identify asthma attacks earlier, guides them through the steps of using the inhaler correctly, and sends an alert when the inhaler is used. The attachment was created with a 3D-printed box enclosing an Arduino Nano 33 BLE Sense Rev2, a stethoscope diaphragm, a button, a 0.96” screen, a SIM800L for SMS, and 3 AAA batteries. The device assists users at each stage of inhaler usage, including shaking the inhaler before use; holding their breath after taking a puff; and repeating the steps if a second dose is necessary. The AI model, developed using Edge Impulse, incorporates comparative analysis of 1D vs. 2D convolutional neural networks along with adjustments in learning rates to optimize detection accuracy.  By removing noise during training and augmenting the training dataset using Python libraries, the 1D CNN with a learning rate of 0.003 was the most successful, with a training data accuracy of 92.4% and a testing data accuracy of 86.07%. Additionally, a web application was developed that allows users to customize the device by adding their phone number. This innovative device, powered by an optimized AI model, enables early detection of asthma attacks and helps manage symptoms through proper use of inhalers.