<p>Morse code is a symbolic communication system that encodes letters, numbers, and characters using short and long signals (dots and dashes). Originally developed for telegraphy, it continues to serve as an efficient input method in assistive technologies, especially for individuals with speech or motor impairments due to its minimal physical effort requirement. This paper presents an assistive communication system that integrates Morse code input with ESP32 microcontrollers and Large Language Models (LLMs) to enable intelligent, natural interaction. A capacitive touch sensor connected to the ESP32 captures Morse code signals, which are decoded into text and transmitted to the Raspberry Pi via a USB serial connection. To improve reliability, an intermediate error-correction stage is applied before final language response generation. This stage reduces dot-dash timing ambiguity and partial-token decoding errors using user-calibrated timing thresholds and context-aware text correction. The Raspberry Pi then forwards the corrected text to an LLM for contextual processing and generates a meaningful response, which is sent back to the ESP32 for real-time display on an OLED or LCD module and to phone. Morse code enables low-bandwidth, reliable communication in environments where speech, typing, or high-speed data transfer is difficult or impossible. This system is especially useful in assistive communication for individuals with speech disabilities, paralysis, ALS, or motor impairments, where even minimal finger or touch movement can generate meaningful communication. Experimental evaluation over 1000 trials shows that the complete two-stage Morse-to-LLM pipeline achieves above 99% end-to-end communication accuracy.</p>