This study explores the integration of a medical assistant chatbot, previously developed with a Long Short-Term Memory (LSTM) model to detect 36 different diseases with over 90% accuracy, into a microcontroller environment. Micro-controllers, known for their low power consumption, compact size, and affordability, present an ideal platform for embedding AI functionalities into portable medical devices. The project involves adapting and simplifying the original chatbot’s LSTM model for use with an Arduino ESP32 microcontroller, addressing challenges such as tokenization and model conversion to fit the microcontroller’s limited resources. The implementation ensures natural language processing, symptom recognition, and diagnostic capabilities are maintained in a compact and energy-efficient device. Performance evaluation demonstrates a validation accuracy of 0.9, precision of 0.86, recall of 0.89, and an F1-score of 0.87, indicating robust and reliable performance. This integration offers a powerful solution for providing real-time medical assistance, particularly in remote or underserved areas with limited access to healthcare facilities.