Define the claim
Who is it about? What intervention, comparison, outcome, dose, and time period are actually being discussed?
Learn
Nutrition questions are rarely solved by collecting more confident opinions. They are solved by asking a narrower question, weighing the right evidence, and translating it without sanding off the uncertainty.
Our evidence standard
We separate evidence quality from presentation quality. A polished chart does not upgrade weak data, and a complicated explanation does not automatically make an answer more honest.
Who is it about? What intervention, comparison, outcome, dose, and time period are actually being discussed?
Study design, consistency, directness, bias, and effect size matter more than a dramatic headline.
We state what is known, what is estimated, and what could reasonably change the conclusion.
The recommendation should be no more certain—or burdensome—than the evidence warrants.
Evidence in the tools
A specific record or body of evidence has been matched and reviewed.
Supports the strongest, narrowest claim the source actually establishes.
A dated value from official packaging or a manufacturer product page.
Useful for that product and formulation—not proof of independent analysis.
A transparent approximation used to support a practical decision.
Useful for planning when the uncertainty is visible and proportionate.
The available source does not support a defensible value.
Shown as unknown. It is never quietly converted to zero.
“We do not know” is a valid finding. Zero is a number, not a placeholder.
Evidence notes
Each note starts with one narrow claim, shows the source boundary, and ends with a practical way to think—not a dramatic verdict.
Why a useful tool needs assumptions, meals, tracking, and feedback—not merely a calorie number.
Read the noteA decision process that integrates research, professional judgment, and the person’s context.
Read the noteWhy missing nutrition data must stay visibly missing until a source supports a value.
Read the noteSee the standard in use