Time Travel
OperationalWhat was this day like? And the same day in every other year?
Renovatio reads the records you already keep — habits, body, money, calendar, place — and shows any day, week, month or year with all of them side by side. It computes on your phone and it keeps everything there.
It is not a tracker. Nothing is entered twice and nothing is estimated. Every figure on every screen comes from a record that already existed: a Google Sheet kept daily since 2018, a Splitwise export, a portfolio file, a wearable, a Timeline export. The app's job is the join, the arithmetic and the honesty.
Every number below is moved and every quoted log entry is replaced. Nothing here is a real record.
C.A.R.L.O.S., shown at larger width — light mode.
Five read the joined record. The sixth reads the other five.
What was this day like? And the same day in every other year?
How is the body doing — HRV, sleep, training load, composition?
Six daily habits, streaks, and the written record.
Household and business spend, by month and category.
What the money actually earned, and what the next 15 years look like.
The day, week, month and year pulled together — what changed, what stopped, what the record supports saying.
The summaries are written by Apple's on-device foundation model. No account, no API key, no request leaves the phone. There is no server, and there is no code in the app that sends personal data anywhere — the only network path it has is read-only, to a compiled-in list of hosts, and it is switched off until you turn it on.
The model is handed findings a script already computed and is not allowed to add to them. Every numeral it writes is checked back against the input before the sentence reaches the screen; one it cannot account for and the whole briefing is discarded, with the reason shown. A summary you cannot check is a summary you cannot use.
Not as a zero, not smoothed over. If spend has not been fetched for six days, the app says the spend record is six days behind, and does not draw a flat line.
Every average carries the number of days that produced it. Every comparison carries the number on each side. A correlation across three hundred days and one across thirty are not the same claim, and the app never lets them look alike.
Sixteen of them, and all sixteen are reported — the ones that found nothing too. Confidence intervals are widened for the number of questions asked. Anything else is a fishing expedition.
When a habit stops for longer than usual, the app shows the gap, how unusual it is against every other gap on record, and what was written in the log that day. It puts the two next to each other. It does not tell you one caused the other.