What we can do — and what we cannot

Body composition measurements are inherently noisy. Weight fluctuates by several pounds day to day due to hydration, sodium intake, digestion, glycogen stores, and hormonal cycles — none of which reflect changes in fat or muscle.

Body fat percentage from consumer scales uses bioelectrical impedance, which estimates rather than measures directly. Results can shift based on hydration status and time of day.

Waist measurements tend to be more stable than weight, but even they vary based on exact measurement point, posture, and time since last meal.

Our approach: we surface the trend line, not the daily reading. We encourage consistent measurement conditions (same time of day, same scale, same measurement method) to reduce noise. We never present a single data point as meaningful on its own.

BMI is a ratio of weight to height squared. It was designed as a population-level screening tool, not an individual health assessment.

BMI cannot distinguish between muscle and fat. A heavily muscled athlete and a sedentary individual can share the same BMI with very different body compositions and health profiles.

BMI does not account for fat distribution, which matters: visceral fat around the waist carries different metabolic implications than subcutaneous fat elsewhere.

Our approach: we include BMI because it is widely used and sometimes required by healthcare providers, but we display it alongside waist circumference and body fat percentage so it is never the sole indicator. We explicitly flag its limitations in the app.

A single reading tells you almost nothing. Two readings tell you slightly more. A trend across weeks or months tells you what is actually happening.

We use rolling averages and trend smoothing to filter out daily noise. This means your dashboard may show a different "current value" than what you logged this morning — that is intentional.

Trends can be misleading if measurement frequency is too low or too irregular. We encourage regular, consistent logging but do not require it.

Our approach: we prioritize direction over magnitude. Whether your waist is declining by 0.3 or 0.5 inches per month matters less than the fact that it is declining. We highlight direction first and specifics second.

Nutrition databases are estimates. Food labels can vary from actual nutrient content by 10–20%. Home cooking introduces further variability in portion sizes and ingredient brands.

Detailed Nutrition mode relies on user-entered data, which is only as accurate as the person entering it. People consistently underestimate intake, especially of fats and calorie-dense foods.

Simple Habits mode intentionally avoids precise calorie tracking. It trades precision for consistency — and for many people, consistency is the bigger lever.

Our approach: we present nutrition data as directional, not exact. We use rolling weekly averages rather than daily totals to smooth out logging irregularity. We never claim calorie counts are precise to the gram.

Ask BodyClarity is an AI assistant that reads your tracked data and explains patterns. It does not have medical training, it cannot see data you have not logged, and it can misinterpret context.

AI models can produce confident-sounding answers that are wrong. They may identify a correlation that is coincidental or miss a factor that a healthcare provider would catch.

The assistant is only as good as the data you provide. Sparse or inconsistent logging limits what it can meaningfully analyze.

Our approach: every AI response includes the underlying data it references, so you can verify the numbers yourself. We never present AI output as medical advice. We include disclaimers on every response. You are always the decision-maker.