Food & symptom analysis
Stop guessing which food is the problem.
Keep a simple diary of what you eat and how you feel. This app runs the statistics a spreadsheet cannot — including which shared compound sits behind seemingly unrelated foods.
Runs entirely in your browser. No account, no upload, no tracking.
A food diary alone rarely gives you the answer
You write down two months of meals, stare at it, and see nothing. The reason is that the culprit usually is not a food — it is something inside several foods. Bread, pasta, onion and garlic look unrelated until you notice all four are loaded with fructans. Milk and hard cheese behave differently because one is mostly lactose and the other is mostly casein. Finding that requires arithmetic no one does by eye.
How it works
- 1
Log it however you like
Type a sentence, pick from a categorised catalog, or fill in a month grid day by day. Symptoms are optional and attach to whichever meal you want.
- 2
Every food becomes a variable
Each food is matched to a catalog entry carrying its compound profile — lactose, gluten, fructans, histamine, capsaicin, and two dozen more.
- 3
Real statistics, honestly reported
For every food and every compound the app builds a contingency table, runs an exact test, and corrects for the fact that it tested dozens of candidates at once.
- 4
An expert reads the numbers
The AI layer never invents figures. It receives the finished statistics and explains the biochemistry that connects them, then proposes an elimination plan you can actually follow.
What you get
Four ways to log
Free text, guided catalog, quick add, or a whole-month grid. Whichever one you will actually keep doing.
Compound-level analysis
The part that matters. Foods are proxies; compounds are the thing you can actually avoid.
Multiple-testing correction
Test fifty foods at p<0.05 and two will look guilty by pure chance. Benjamini-Hochberg keeps that in check.
Confound detection
If you never eat bread without butter, the app says so instead of blaming one of them at random.
Bilingual and offline-first
English and Spanish throughout, including food-name matching. Your data never leaves the browser.
Export anytime
One JSON file with everything. Import it back, move it to another browser, or hand it to your dietitian.
What this app will not do
This is a hypothesis generator, not a diagnostic tool. It is designed to be sceptical of itself.
- It will never tell you that you have an intolerance. It tells you what is worth testing next.
- Every association ships with its uncertainty, its sample size and a corrected q-value.
- Correlation is not causation, and eating patterns are deeply confounded. The app flags that instead of hiding it.
- If you record an alarm symptom — blood, weight loss, night pain, trouble swallowing — it stops analysing and tells you to see a doctor.
- Suspected coeliac disease must be tested by a doctor before you remove gluten, because removing it first invalidates the test.
Your data stays yours
Everything is stored in your own browser using IndexedDB. There is no account system, no server database and no analytics. The only network request the app can make is the optional AI analysis, and that sends computed statistics — never your raw diary.
Ready to find the pattern?
Two weeks of honest logging is usually enough to see something. A month is better.