Log a meal by describing it
Type what you ate the way you would say it to a friend. Aurora AI reads the portion, the cooking method and the sides, then estimates the calories, protein, carbs and fat.
It knows your usuals too. Foods you eat often appear as one-tap chips, and every food you have ever logged is searchable, so a repeat breakfast takes one tap rather than a fresh description.
Every estimate is itemised, so you can check the maths
An AI estimate is worthless if you cannot audit it. Every meal expands into its components, each with its own calories and macros: 2 eggs at 140, toast at 180, and so on.
If a number is wrong, correct it in a sentence. "The chapatis were 125 kcal each" updates the existing entry rather than logging a duplicate, and the breakdown is recalculated.
Exercise raises your budget instead of being ignored
Log a run or a gym session, or paste the burn straight from your watch, and those calories are added back to the day. A bigger dinner after a hard session is not flagged as overeating, because your allowance genuinely went up.
The calorie chart reflects this: each day's column is that day's real budget, so training days visibly rise above your baseline target.
A score, not just a number
Every logged day is scored out of 100: calories inside your adjusted budget, protein hit, movement, and logging consistency. You can see at a glance whether Tuesday was actually good or just quiet.
Over 14 days that becomes a pattern, and the coach tells you what is working, what to fix first, and what your current pace does to your body over months. It is built from your real behaviour, not the plan you set on day one.
Your data stays yours
Export everything as fine-grained CSV whenever you like: every meal, every component, every workout, and a per-day rollup with your net calories against your adjusted target.