Plant Disease Detection using Deep Learning is a powerful Streamlit-based web application designed to identify potato leaf diseases such as Healthy, Early Blight, and Late Blight. Utilizing a trained Convolutional Neural Network (CNN) model built with TensorFlow and Keras, this application offers an end-to-end workflow from data preprocessing to model training, evaluation, and deployment. Key features include the ability to upload leaf images for instant disease predictions, complete with confidence scores. The interactive dashboard provides visualization tools like a Confusion Matrix and performance comparisons between validation and test datasets. The application is easily deployable with Streamlit and uses a lightweight .h5 model format for portability. The dataset is sourced from the PlantVillage dataset, containing approximately 2,152 images split into train, validation, and test sets. The model architecture features multiple Conv2D and MaxPooling layers, followed by Flatten and Dense layers, with a Softmax output for three classes. Data augmentation techniques such as Random Flip and Random Rotation are employed, and the model is optimized using the Adam optimizer with SparseCategoricalCrossentropy loss. Achieving high accuracy rates of 96.88% on validation and 97.84% on test datasets, this application is a robust tool for plant disease detection.