The framework integrates a puzzle engine, a life-logging module for personal memory recall, and a multimodal training interface featuring text, image, and voice capabilities. Cognitive training is one of the few non-pharmacological methods shown to help delay decline, but existing programs are often generic and not very engaging. ReMe aims to make cognitive training more personalized and adaptable while reaching more people at lower cost. Instead of building every new cognitive training chatbot from scratch, researchers can use ReMe to prototype, test, and improve interventions more quickly, speeding up discovery of what works. In initial evaluations, ReMe was well received by users citing strong conversational fluency and moderate difficulty. While not intended for clinical treatment, ReMe provides a valuable tool for exploring AI’s role in supporting cognitive health, paving the way for future innovations in personalized digital therapies.