The dangerous disease and ailment known as COVID-19 has caused a global pandemic and immeasurable damage to people all around the globe. Chest X-Rays are being used to identify COVID-19 in current times due to their low cost and efficiency. In this paper, we developed Convolutional Neural Network models to detect COVID-19 from Chest X-Ray images. CNN models from the Keras library such as VGG16, VGG19, Xception, ResNet101, ResNet152, Inception, InceptionResNet, MobileNetV2, DenseNet201, NASNetLarge, and EfficientNetB3 have been used to perform experimentations. The CNN models used ImageNet as its pre-trained weights for transfer learning. Additionally, a multi-layered self-designed model has been implemented as well to see the performance. A comparative analysis has been completed in order to find the best-performing CNN model for COVID-19 detection in Chest X-Ray images. From the experiments, we found that the proposed CNN gave the best results. Additionally, it has been observed that MobileNetV2, Inception, ResNet101, and VGG16 give the highest accuracy over 99%, while the lowest accuracy is found by EfficientNetB3 at only around 50%. The self-designed multi-layered model gives a training accuracy of 97.22% and a validation accuracy of 96.42%. A significant increase in accuracy and excellent performance has been seen from the CNN models and the proposed framework.