—
NET KCAL
Intake
—
Burned
—
BMR (rest)
—
NET = Intake − Burned − BMR (your resting burn). A negative net means a calorie deficit.
Weight & circumference over time
weight (left)
circumference (right)
NET = Intake − Burned − BMR (your resting burn). A negative net means a calorie deficit.
Signed in as —. Profile drives your BMR.
Food logged to a meal is timestamped at this time of day.
Log starts, dose changes and stops. Doses change roughly monthly, so this is an event log — not something to enter daily. The model needs it because GLP-1s change how calories translate to weight, and independently affect drinking, which would otherwise corrupt your per-drink numbers. Nothing here is medical advice.
Every AI identify call is logged below. Cost is calculated from the token counts the API returns — the API never returns a dollar figure — using the published per-model rates, so treat it as a close estimate rather than a bill.
The password you enter once. If you see “Unauthorized”, set it here.
Anything you add now is a plan — it shapes tomorrow's projection but doesn't count as eaten until you confirm it.
Tap to plan it for today. Long-press or use ✕ to delete.
Searches your saved foods and the USDA database. No match? Ask the AI to estimate it — you can adjust the amount before logging. The Photo tab estimates the portion from a picture.
Tap an item to edit it, or ✕ to remove it.
Tap an item to edit it, or ✕ to remove it.
Enter either or both — weight only, waist only, or both.
Tap a measurement to edit it, or ✕ to remove it.
Weigh a real plate on a kitchen scale, then photograph it the normal way. This teaches the app the correction factor for that food's shape — after enough runs per category, you won't need to weigh anymore.
Not found by search? Describe it and AI will estimate (⚠ costs ~1¢ the first time):
Pick a food or drink to see how the model's predicted weight impact of one typical serving has drifted from the 3500-kcal/lb baseline as it learns from your weigh-ins.