About

Day So Far is built, run and paid for by Nikolay Lyutov, an independent developer established in the European Union. There is no company behind it, no investors and no team — the same person writes the code, answers support@daysofar.com, and is the data controller named in clause 1 of the privacy policy.

That is worth stating plainly rather than hiding behind a “we”. It sets expectations in both directions: replies come from a person and sometimes take a day, and there is nobody to escalate to — but there is also nobody whose job depends on keeping you subscribed, and no growth target that a darker design pattern would help hit.

Every calorie tracker asks you to search a database. You type “chicken breast”, you get forty rows that differ in ways you cannot see, and you pick one. Do that three times a day and the friction is not the arithmetic — it is the picking. Most people stop within a fortnight, and not because they stopped caring what they ate.

So this one has no search box. You describe the meal the way you would tell someone about it, and a language model works out what you probably meant and looks the nutrition up. Photograph the plate, scan a barcode, or say “my usual” instead, if that is faster. The point is the same in all four cases: no list to choose from.

The app produces estimates. It is not a laboratory, not a dietitian and not a medical device, and the numbers it gives you carry real error — which is measured rather than waved away. The figures are on the accuracy page, including the ones that do not flatter it.

Two things follow from that, and both are design decisions rather than disclaimers:

• A guess is shown as a guess. Where the app has had to assume a portion, it says so rather than presenting an assumption as a measurement.
• Trends are worth more than any single day. A calorie target that adapts to what your weight actually does — see how it works — survives individual estimates being wrong in a way that a fixed formula does not.

By subscription, and only by subscription. There is no advertising, nothing is sold to anybody, and no analytics product is watching you use it — the privacy policy names every third party that ever receives anything and why. The journal itself is free; the parts that cost money to run are what the paid tiers pay for.

The app is open source and can be run on your own server. If you are using somebody else’s installation, the privacy policy here does not describe it — theirs does, and their operator is the controller, not me.