Stroke is a critical global health issue, with minor strokes often going undetected due to subtle symptoms. These minor events are key indicators of potential major cerebrovascular issues and, if untreated, can lead to severe health deterioration. This paper presents StrokeAlert, a mobile application for the early detection of minor strokes using the FAST criteria (Face, Arms, Speech, Time). The app employs image analysis, speech recognition and arm weakness detection technologies to detect stroke symptoms in real time using deep learning algorithms. Testing demonstrated an 83.84% accuracy in facial drooping detection, 96.33% in speech difficulty evaluation and 88% in arm weakness detection. StrokeAlert offers an accessible, user-friendly solution that could significantly enhance early intervention and reduce stroke related complications.