We Build a Canonical Recipe
Each dish page points to a Bonutivo homemade recipe archetype — not a random blog post. Regional home style, cooking method, water/oil, and servings are fixed for that published version.
Every nutrition value on Bonutivo is built from a transparent, reproducible process — not estimated by AI.
Unlike sites that invent macros for a dish name, we start from measured ingredients, a fixed homemade recipe version, cooking science, and pinned reference data. The number you see can be explained step by step.
From ingredients to trustworthy nutrition on your plate.
Each dish page points to a Bonutivo homemade recipe archetype — not a random blog post. Regional home style, cooking method, water/oil, and servings are fixed for that published version.
Kitchen names (rava, sooji, suji) resolve to one ingredient identity with IFCT-oriented composition per 100 g raw — so aliases never silently change the macros.
Raw weight is not the same as what lands on the plate. Yield and retention factors adjust mass and nutrients for boiling, tempering, shallow frying, and similar home methods.
Calories, protein, carbs, fat, fibre, sodium, and sugar are computed from the canonical recipe lines plus pinned calculator rules (including Atwater energy checks).
Ingredient completeness, unit conversions, oil/water consistency, and batch checks must pass before a result can be published to the catalog.
Published snapshots cite a calculator release and reference dataset package. When science or a recipe improves, we ship a new version instead of silently editing the old one.
From the dish page you can follow the chain that produced the macros:
Two independent assessments keep our system honest and transparent.
★★★★★
★★★★★
A beautiful recipe with a broken calculation is useless. A perfect calculator on a vague recipe is also useless. Dish pages show both Recipe Quality and Calculation Quality so you can see where trust comes from.
Nutrition science and home cooking both move — our catalog is built to update without hiding the trail.
New retention research or composition updates become a new reference release, not a silent overwrite.
When a homemade pattern changes, we publish a new canonical version and keep the history.
IFCT / NIN-oriented packages are pinned so old snapshots stay reproducible.
Explainable steps beat black-box AI nutrition guesses for household decisions.
Give every family explainable, reproducible, and auditable nutrition for the homemade dishes they actually cook — so choosing better food feels calm, not mysterious.