Emotion recognition aims to create AI systems that figure out different emotional states. For years, experts from fields like computer science, anthropology, and ethics have dug into this area working hard to come up with new models and methods. The main way to uncover hidden emotions is to spot facial expressions. Using convolutional neural networks (CNN), various models have been trained to do this. This tech has uses in many areas such as health, safety, and traffic. In this project, we show a CNN-based model that can spot seven basic emotions: anger, disgust, fear, happiness, sadness, surprise, and neutral. We used the FER-2013 dataset split into training, testing, and validation sets. After cleaning up the data, we built a CNN model. When we tested it against the dataset; it gave good results with decent accuracy.