On food tracking
Why We Built a Food Tracker You Barely Have to Touch
Take a photo, say what you ate, or type a sentence. That is your whole job. A note on why every food log has holes in the same places, and what changes when logging stops being work.
I've tracked my food for years. I started because I had no real sense of how much I was eating, and it turns out you can't guess your way there. Calories and macros aren't intuitive. You measure for a while, and then your estimates start being worth something. The habit has taught me more about my body than anything else I've tried.
Then one day I looked at the record instead of the numbers.
61% complete. 9 days with nothing recorded at all. The filled days are the meals eaten a hundred times before — same chicken, same rice, logged from muscle memory. The blanks are a trip, two dinners out, and one night improvised from whatever was in the fridge.
I. The holes were never random
My log was flawless at the meals I already understood, and empty for every meal I didn't. That is not a discipline problem, and reading it as one is how the entire category gets its users wrong.
Getting an unfamiliar meal right meant breaking a plate into a dozen searchable parts, one search at a time, then judging between four near-identical entries for each one. Three or four times a day. Do that long enough and you find the exact point where a habit becomes a chore, and then becomes something you used to do.
II. What tedium actually costs
Every tracker on the market sells accuracy. Almost none sell ease, because ease sounds like a compromise. Here is what actually happens when logging takes real work: you don't quit on day one. You start cutting corners.
- You log the chicken and skip the olive oil it was cooked in, because the oil is one more search for 120 calories.
- You take the first entry that looks close instead of the right one.
- You round the rice.
- You log Monday through Thursday and reconstruct Friday from memory on Sunday.
The app is still accurate. Your data isn't. That's the part nobody says out loud: accuracy is not a property an app has. It is something a person produces, meal after meal, and people only produce it when it is easy.
Every extra step is charged to your data. We are not trading accuracy for convenience; we are building the only version of accuracy that survives contact with a real week.
III. Tell it what you know
Point your camera at the plate. Say what you ate. Type a sentence. One step, and your part is over. You do not wait on it either: the analysis runs whether you are watching or not, so you can put the phone down and go back to the table, and your log is there when you next look.
You usually know more than a photograph shows, and every bit of it should count for something.
| What you give it | Detail | What you get |
|---|---|---|
| A photograph of the plate | Estimate | A good estimate |
| A sentence, spoken“Chicken thigh, 180 grams, jasmine rice, two cups” | Detailed | Near-exact |
| A barcode | Exact | Exact, off the label |
Precision was never proportional to effort. The resolution is yours to set, and raising it costs you words, not another trip through a search field. Underneath sits verified nutrition data, barcodes, your saved meals and a food database that improves as it is used — none of which you ever have to go digging through yourself.
IV. Two kinds of people track food
Some of us eat differently every day. Travel, restaurants, whatever was in the fridge, the same meal made three slightly different ways in one week. For us, the effort per meal is the entire ballgame, because every variation is otherwise a small research project.
And some of us eat close to the same thing for months. If that's you, you shouldn't be re-logging anything: save the meal once, tap it, done. When your goal changes — cutting to maintaining to gaining — the portions move with it, instead of making you rebuild the day around new numbers by hand.
Most apps pick one of these people. We think they want the same thing, which is for this to be simple.
V. No catalogue holds every meal
That isn't a knock on databases — it is what catalogues are. They are built from packaged goods and standardised entries, they are very good at those, and we lean on them hard. They were also never going to contain your grandmother's cooking, or a dish from a market stall in a city you are visiting for four days.
A model does not need a catalogue entry to recognise food. It can look at a dish it has never been formally introduced to, give you a real answer, and let you sharpen it with a sentence. That is the difference between a tracker that works where you live and one that works wherever you go. We intend to be the second kind: on your food, in your language.
VI. Before you eat it
For some people the calories are not the urgent part. You need to know whether there is dairy in the dressing, whether the sauce contains the one ingredient that costs you a night's sleep, whether this plate undoes months of getting your symptoms under control.
Work from a curated trigger list or build your own. Log something that contains it and you get told — before you eat it, not at two in the morning, when the information is useless.
VII. What it costs
| Free | $0 | Diary, daily goals, saved meals, barcodes, three AI logs a week. |
|---|---|---|
| Plus | $6 / month | Unlimited photo, voice and sentence logging. Trigger warnings. Meals that rescale with your goal. |
| Plus, annual | $48 / year | The same, at two months off. Waitlist members pay $36 for as long as they stay. |
No advertising. We do not sell what you eat, and no tier holds your own history hostage.
VIII. This is step one
Here is the uncomfortable part of nutrition: most people who genuinely understand what they eat bought that understanding. A coach, a nutritionist, a food scale and six hours a week. Everyone else is guessing, and has been told, more or less constantly, that guessing is a personal failing rather than a tooling failure.
It isn't. Nobody is born estimating macros. You learn it by measuring, and measuring has always cost enough effort that most people never got the education at all. Take that cost to almost nothing and something changes — not just for the guy cutting for summer, but for the person managing a condition, the person who has never tracked anything and is quietly curious, the person who needs to know for one season and then wants their attention back.
Understanding your own body stops being a luxury good. That is the world we are building toward. The app is where it starts.