The method. The instrument. The reading.

01The setup — why n=1 matters

Population data tells you what is likely true on average. It will not tell you what is true for you. The studies and protocols you've already read are very good at the first job; they are silent on the second.

Most people who arrive at n1lab arrive after years of trying the protocols. They have been told what their body should do. They are here because what their body actually does turned out to be information no study was looking for.

The aim is modest. Record carefully. Wait. Repeat the trial. Hold every reading lightly until a second reading agrees with it.

That last sentence is the whole instruction set. Everything in the app is in service of keeping that loop clean — making sure the next observation sits next to the last one, scaled the same way, on the same baseline, so a pattern has the chance to be a pattern instead of a feeling.

02The instrument — what the app records

n1lab records six things. Each one is a separate surface; none of them combine into a single number, because a single number would assert more than the readings support.

1. AM check-in. Sleep, energy, mood, anything notable from overnight. Logged inside a four-hour morning window; outside it, the field closes.
2. PM check-in. End-of-day burden across symptom domains. When both check-ins are recorded, the day closes.
3. Symptom burden, eight domains. Cognitive, mood, energy, gut, sleep, pain, cravings, autonomic. Each scored 0–10. A skipped domain is recorded as null, not zero.
4. 24h food trial. A reintroduction observation with three intra-day windows — 0–2h, 2–6h, 6–24h — and a baseline drawn from the preceding fourteen days.
5. Weight. Plotted against a chosen baseline. No goal numbers. No setpoints. The trend is the only thing the chart shows.
6. Weekly review. A short Sunday-evening pass over the week's logs. The app surfaces what it noticed; you decide which observations to keep.

03The protocol — daily readings

Two check-ins a day, a few seconds each. The check-ins are not journal entries; they are readings. The format is fixed so that the readings are comparable to each other.

03.1The morning reading

Sleep duration. Sleep quality. Energy on waking. Mood. A free-text field for anything notable from overnight. The free-text field is short by design — long entries reward elaboration, and elaboration is the enemy of consistency across weeks.

03.2The evening reading

Eight domain scores, 0–10. The slider for each domain is anchored to your recent range, not a universal scale, so a score of 6 means the same thing it meant last Tuesday. The app does not show you a "normal" — there isn't one.

04The trial — a 24h observation

A trial is a single 24-hour window opened around the reintroduction of one food. The trial captures the food, the dose, the start time, the confounders you noted at the start, and the symptom load in each of the three intra-day windows.

The close-out screen lays this out in a single view: a seven-day outer chart centered on the trial day, a three-bar intra-day strip beneath it, and one row per relevant symptom. The trial does not output a verdict. You confirm the interpretation. The app records the observation, the evidence count, and the confidence.

05The baseline — what counts as elevated

Every chart in n1lab is drawn over a band, not a line. The band is your fourteen-day rolling mean plus or minus one standard deviation. A reading inside the band is unremarkable; a reading outside it is the app's narrowest definition of "something happened."

The band moves. As your baseline shifts — better sleep, fewer flares, a new medication, a season change — the band shifts with it. What counted as elevated in April will not necessarily count as elevated in July. This is on purpose. The baseline is your recent body, not your aspirational body.

06Confidence — what the bars mean

Every recorded observation carries a confidence — low, moderate, or high — and an agreement count. The three bars on a confidence line are not a score; they are a literal count of how many of the tracked symptoms support the interpretation you chose.

One bar is one observation in agreement. Three bars is total agreement across the tracked symptoms. The number that follows — "3 of 4" — is the agreement count out of the symptoms that had data. You should interpret a moderate-confidence reading as a hypothesis worth re-testing, not a fact worth acting on.

The app does not let you record a "very high" confidence from a single trial. The vocabulary tops out at "high," and the agreement count tops out at four. To get past that, you repeat.

07When to repeat. When to stop.

Repeat when the first trial returned a reaction and you want to know whether it is real. The recommended interval is four to six weeks; long enough for the immediate signal to clear, short enough that other variables in your life have not all shifted.

Repeat when the first trial returned tolerated and your gut tells you otherwise. The point of n=1 is precisely to take that gut feeling seriously by asking it to show its work twice.

Stop when two repeat trials have agreed and the food, dose, and confounders were comparable. Two agreements is the working ceiling for n1lab. Beyond that, you are no longer observing; you are confirming.

Stop when the trial is no longer safe — a flare bad enough to require care, a confounder you cannot remove, a context where the question stops being worth the answer. The app does not gate this for you. You are the instrument's operator.

08The limits — what this is not

n1lab is a recording surface. It is not a clinical instrument, and it does not claim to be one. The patterns it surfaces are suggestive at best; they are intended to be brought into a conversation with a clinician, not to replace one.

If a trial returns a serious reaction, stop the trial. If a symptom changes in a way that frightens you, the right next step is care, not another reading.

Everything else, n1lab is glad to record.