Farmers are tasked with making numerous critical decisions each month, where the identification of crop diseases poses significant challenges due to its potential economic ramifications and effects on food supply. Early detection of diseases is crucial for mitigating damages and preventing their spread. This work proposes a GAN-based model for detecting tomato plant diseases, achieving an overall accuracy of 91% and a weighted F1-score of 0.87. The model extends conventional GAN functionality to multi-class classification, allowing the discriminator to identify whether tomato plants are healthy or diseased and, if diseased, classify the type of disease. With efficient training and robust generalization to unseen data, the model has been deployed as an API for real-time disease detection. Future work will focus on expanding the dataset diversity and integrating multi-language support to improve accessibility. We will first explore the literature survey to understand the existing work and foundational concepts related to our project. Following that, we will discuss the project architecture in Sect. 2, dive into the structure of the GAN model and its discriminator architecture in Sect. 4, and then move on to the result analysis of the model and the comparisons in Sect. 5. The report will conclude with a summary of our findings, and the references will be provided at the end.