The failure mode of every health app is the same. They want input. Type in your meal. Type in your sets. Type in your weight. The work of measuring your life ends up taking longer than the life you are trying to measure. Most people quit by week three.
The Health Hub in AssistantApp is built around the assumption that the typing is the problem. Take a photo of the plate. The food-scan handler uploads the image and a vision model returns calories and macros — protein, carbs, fat — along with portion estimates. The day’s nutrition log fills in without a keyboard. It costs a few credits. It does not cost a habit.
Workouts get the same treatment. A reusable strength template pulls up the lifts you did last Tuesday and lets you plus-or-minus the weight in a tap. Cardio templates do duration and distance. Calorie burn calculates itself out of that. The HealthKit cache pulls the Watch’s active calories and steps in the background, so a walk at lunch counts whether or not you remember to log it.
Body metrics live in their own corner — weight, body composition, progress photos — visible only to you. The measurements stay private by default. There is no community feed. The app does not know your friends.
Once a week, an AI coaching pass writes a short report. Not a leaderboard. Not a streak number. A note: protein hit goal four days out of seven, the workouts that produced the best perceived energy were the morning lifts, sleep on training days was an hour shorter than rest days. The user can ask follow-ups in plain text. The model does not give medical advice; it gives the kind of pattern you would notice yourself if you had an hour on Sunday to think about your week.