Why food is the hardest thing to track

Everything here is delayed, dose-dependent and confounded

Which is exactly the combination human memory handles worst, and exactly the combination a timestamped log handles well.

The lag defeats memory

A non-allergic reaction can arrive six hours or three days after the food. By then you've eaten roughly ninety other things, and the one you blame is the one you happen to remember rather than the one responsible.

Dose matters, so yes-or-no fails

"I ate it and I was fine" proves nothing if the amount was small. Most intolerances are thresholds, not switches — which is the difference between never eating something again and eating it in normal quantities.

Everything else moves too

Stress, sleep and the menstrual cycle shift gut and skin symptoms as much as food does. Track food alone and you will confidently blame a food, narrow your diet, and not get better.

Where this actually helps

Food diaries fail for one boring reason

They ask for more work than anyone will do for more than nine days. Weigh it, categorise it, find it in a database of 40,000 products. Meanwhile the useful entry was always going to be one sentence long.

Say what you ate and how you felt. The precision that matters here is timing, not grams — knowing a symptom started twenty hours after a meal is what makes a reintroduction interpretable, and a gram count adds nothing to that.

  • Log a meal or a symptom in one line, typed or dictated
  • Backfill by saying "yesterday" or "since Monday"
  • Reintroduction ladders kept on schedule rather than improvised
  • Dose recorded, because most of this is a threshold rather than a switch
Before a dietitian appointment Chat

Dietitian on Tuesday. What did the dairy and wheat reintroductions actually show?

Called 3 tools

  • get_challenges(protocol: "elimination")
  • compare_challenges(foods: ["dairy", "wheat"])
  • export_data(range: "protocol", format: "pdf")

Dairy: two trials, bloating at 18–22 hours in both, cleared inside 48. Wheat: two trials, nothing logged at any interval. Background stress was high in one dairy week and low in the other, and the result held either way. Exported both as a one-page summary.

Timings and counts from your own entries. What it means, and what to do about it, is Tuesday's conversation.

Questions

The ones worth asking

Can I use this to work out if I'm allergic to something?

No, and please don't try. Allergy is a clinical diagnosis with actual tests behind it, and the home version — eliminate it, then eat it and see — is an oral food challenge, which is done in a clinic with adrenaline present because it can go wrong in minutes. This is for managing an allergy you already have a diagnosis for, and for the non-allergic reactions where tracking is genuinely the method.

What's the difference between allergy and intolerance?

An allergy is an immune response, usually fast, and can be life-threatening. An intolerance is digestive or metabolic — slower, dose-dependent, unpleasant rather than dangerous. They need completely different handling, which is why they have separate pages here. Anything involving breathing, throat or tongue swelling, or faintness is the first kind and is an emergency.

Should I go gluten-free to see if it helps?

Not before you've been tested for celiac disease. The tests look for your immune response to gluten, so removing it first makes them come back negative whether or not you have it — and getting a real answer afterwards means months of deliberately eating something that makes you ill. Test first, then eliminate.

Do those at-home intolerance tests work?

No. IgG food panels, hair analysis and applied kinesiology aren't supported by evidence and reliably return long lists of foods you're perfectly fine with. The thing that actually works is a structured elimination and reintroduction with a dietitian, which is what this is built to support.

Will the AI tell me what I'm reacting to?

It'll tell you what your log shows — that symptoms followed dairy at 18 to 22 hours in both trials, that wheat produced nothing, that the effect held even in the low-stress week. It reports counts and timings. Turning that into a conclusion, and into a diet, is a job for you and a dietitian.

Isn't tracking every meal a bit unhealthy?

It can be, and that's worth taking seriously. Elimination protocols are meant to be temporary and to end with a wider diet than you started with, not a narrower one. If logging food is making your relationship with eating worse rather than better, that's a good reason to stop and talk to someone — and a better reason than any pattern in the data is to keep going.

Do I have to weigh and log everything?

No. Rough amounts are enough for almost all of this, because what makes a reintroduction interpretable is the timing rather than the gram count. The one place amount matters is dose thresholds, and "a latte and a yoghurt" is a good enough dose for that.

Early access

Start with the one you're already chasing

Tell us what you've been trying to pin down and we'll set it up with you — recording the thing that makes the answer interpretable, at a cadence that survives past week two.

  • Founding accounts lock their price for life
  • No card, no trial clock, no notifications designed to make you feel behind
  • Export everything you've logged whenever you want

No card required. Unsubscribe any time.

Find the pattern without living in a food diary

Founding pricing is locked for accounts opened during early access.

Founding accounts lock their price for life. No card required.