Main

Rapid advancements in large language models (LLMs) have driven transformative changes in health. Current LLMs can support various clinical tasks, such as medical documentation 3 , diagnosis 4 , decision support 5 and drug discovery 6 , 7 , as well as consumer health tools, including virtual health assistants 8 , symptom checkers 9 , digital coaches and advisors 10 , 11 , and clinical examination dialogues 12 .

Wearables add continuous activity, sleep, heart rate variability and other physiological signals 13 , 14 , providing longitudinal context for health support. Recent systems connect these data to LLMs 1 , 2 , 8 , 15 , 16 , 17 , but numerical time-series reasoning remains unreliable 18 . Previous systems either cover narrow fitness or sleep questions without deep numerical analysis 8 , or address broader numerical queries without evaluating health expertise and interactive coaching 2 . A comprehensive system must combine these capabilities while maintaining a coherent interaction 19 .…