New to raw feeding and not sure where to start?
- You don't need a nutrition degree - the engine does the math
- Every number comes with a plain-language explanation
- Your vet can review the output - it's designed to be shareable
You enter your dog's profile and ingredients. The engine calculates how much of each food to feed, checks the recipe against NRC 2006 nutrient targets, and tells you where the gaps are - with specific food-first fixes before any supplement is suggested. It works differently for a 5 kg Chihuahua and a 40 kg Labrador, for a puppy and a senior, for a dog with kidney disease and a healthy adult. One size does not cover all of them.
Three things Raw & Well checks that most calculators miss
Whether your dog is actually absorbing what's in the bowl
Zinc in raw meat has roughly 35% bioavailability. If the engine counted what's in the food rather than what crosses into the bloodstream, every zinc target would be wrong by about 3x. It doesn't. Absorption coefficients are built in.
Whether your bone percentage is actually delivering the right minerals
Chicken necks and beef marrow bones are not the same thing at 10% by weight. The mineral yield per gram differs significantly by species and cut. The engine knows this. The "10% bone" rule does not.
Whether a zinc supplement is blocking copper absorption
Zinc and copper compete for the same intestinal transport proteins. A recipe that looks fine on both minerals separately can still produce copper deficiency when zinc is consistently elevated. The engine tracks both in the same energy-based unit system so the interaction shows up.
What this looks like for a real dog
Bella is a 4-year-old female Vizsla, 22 kg, moderately active, intact. Her owner had been feeding chicken necks as the primary bone source at 10% by weight - a structurally sound recipe on paper.
10% bone, 80% muscle meat, 10% organ. Clean ratios. No obvious problems.
Calcium sitting at 0.8% dry matter. NRC's recommended allowance for a dog Bella's size: closer to 1.0%. Phosphorus elevated from organ content. Calcium-to-phosphorus ratio: 0.85:1 - below the safe floor of 1.1:1. She'd been eating that way for months.
Bone source switched from chicken necks to a higher-calcium-density cut. Organ fraction trimmed slightly. New ratio: 1.2:1. Same percentage framework. Same feeding structure. Different mineral outcome.
Nothing. "10% bone" looked fine either way.
Why calculators disagree - and which one to trust
The gap between Raw & Well's output and a basic calculator isn't a rounding difference. It comes from a fundamental choice about what kind of math to run.
Generic calculators
- Linear scaling: body weight – a fixed multiplier
- Ignores the metabolic exponent (\(kg^{0.75}\))
- No breed-specific or life-stage adjustments
- Gets further from correct at the size extremes
AI chatbots
- Predicts text - does not run a calculation
- Blends conflicting sources into a single confident-sounding answer
- No hard-coded safety ceilings
- Cannot track a dog's specific profile over time
Raw & Well engine
- Allometric scaling (\(kg^{0.75}\)) - Kleiber's Law
- Runs a defined 6-step sequence for every calculation
- Life-stage and breed context applied as labeled assumptions
- NRC 2006 Safe Upper Limits enforced as hard ceilings
Ready to see this in action for your dog?
7-day free trial – No card required – Cancel anytime
Start freeThe 6-step calculation - what happens when you press Calculate
Every recipe runs this sequence. The steps are fixed. The inputs are yours.
Step 1 - Biological baseline
Your dog's 8-point profile (weight, life stage, breed, neuter status, body condition, activity level, feeding model, clinical flags) goes through Kleiber's Law: \(\text{RER} = 70 \times (\text{BW}_{\text{kg}})^{0.75}\). This gives the resting energy requirement. Activity, life-stage, and growth multipliers then produce the day-to-day energy requirement (DER) - adults scale resting energy by activity, while growing puppies scale the NRC maintenance amount by the Table 15-2 growth curve.
Step 2 - Safety ceilings applied
Before a single ingredient is touched, the engine locks in NRC Safe Upper Limits as hard ceilings. Where a veterinarian has confirmed a diagnosis (CKD staging, copper-sensitive breed), separate clinical policies are applied and labeled as such - distinct from NRC table values.
Step 3 - Gram portions calculated
You pick ingredients from five categories (muscle meat, raw meaty bone, liver, secreting organ, optional produce). The engine calculates personalized gram amounts - meeting NRC targets from Step 1, respecting the ceilings from Step 2. Same ingredients, different dogs: a 5 kg Chihuahua and a 35 kg Labrador selecting identical ingredient types receive very different amounts.
Step 4 - Bioavailability adjustment
What's in the food is not what the dog absorbs. Zinc in raw meat absorbs at roughly 35%. The engine applies bioavailability coefficients before checking nutrient totals. If the coefficient for an ingredient is uncertain, the engine treats the yield as an estimate and keeps that visible - rather than printing a precise number it can't support.
Step 5 - Gaps identified, food-first fixes offered
Every absorbed nutrient is cross-referenced against NRC Recommended Allowances. Where a gap exists, the engine checks whether a whole-food fix is practical given the dog's caloric budget (the Volumetric Ceiling). If adding more food would push calories too high, it routes to a measured supplement recommendation instead. Supplements are not the first answer.
Step 6 - Symptom correlation overlay
When you log an observation (stool quality, energy level, coat changes), the engine maps it against that day's exact macro ratios - bone percentage, meat percentage, organ percentage. If a loose stool flag coincides with a high-organ day or a new protein introduction, the engine surfaces that connection. You provide the observations. The engine provides the map.
Data sources: USDA FoodData Central
A calculation is only as trustworthy as the numbers feeding it. Raw & Well does not estimate ingredient nutrition; it traces it to a government database you can check yourself.
Soft-tissue nutrient values for the ingredient database are sourced directly from USDA FoodData Central, the U.S. government's authoritative food-composition database, drawn from its SR Legacy and Foundation Foods datasets. Each ingredient carries a verified FDC ID and a direct link to its official USDA entry, so any number in your meal plan can be independently verified against the same source used in peer-reviewed veterinary research. This is government-traceability, not a generic “USDA-based” label: Beef Liver is FDC 169451, raw chicken liver is FDC 171060, Eastern wild oyster is FDC 171978, each auditable in one click.
Where USDA does not publish data, including the whole-bone calcium and phosphorus of raw meaty bones (USDA measures only deboned flesh) and 12 exotic ingredients such as kangaroo and eggshell powder, the app flags the data source explicitly rather than estimating. Whole-bone minerals come from established raw-feeding literature, which keeps the calcium-to-phosphorus ratio clinically correct instead of artificially near-zero.
Anatomical bone percentages — the share of each raw meaty bone that is edible bone — are not published by USDA or NRC, which measure nutrient composition rather than carcass structure. These values are estimated from published carcass analyses (Monica Segal, RVT) and raw-feeding reference data, and are labelled as estimates in the app. Actual bone content varies with cut, age, and butchering.
The resulting calcium-to-phosphorus band is verified compliant with both NRC 2006 and FEDIAF 2024: an absolute minimum of 1:1, an adult maximum of 2:1, and a growth ceiling of 1.6:1 (FEDIAF permits up to 1.8:1 for late-growth large breeds; Raw & Well holds the tighter 1.6:1).
Three engines. One platform.
A growing Great Dane puppy, a healthy adult Dalmatian, and a senior Labrador with kidney disease need different math. The engine routes each dog into one of three specialized pipelines based on the 8-point profile. The routing is automatic. The assumptions are visible.
The 8-point clinical profile
Every data point changes the output. Here's what each one does.
The base for \( kg^{0.75} \) scaling. A 10 kg dog and a 20 kg dog don't just eat twice as much - their metabolic relationship is different.
Routes the dog into the correct engine. Growth, adult maintenance, and senior each have different NRC targets and different safety sensitivities.
Applied as optional context modifiers - not as a substitute for energy-based scaling. Breed matters for copper-associated hepatopathy risk, for large-breed growth sensitivity, for toy-breed caloric density concerns.
Neutering causes metabolic shifts that affect caloric needs. The engine accounts for this rather than ignoring it.
Adjusts the baseline for weight management goals. A dog at BCS 7/9 needs a different target than the same dog at BCS 5/9.
Sets the day-to-day energy requirement (DER). A working dog and a couch dog of the same weight have genuinely different needs.
BARF, PMR, or Custom. Sets the architectural ratios the Meal Planner builds around.
Activates safety overrides: IRIS kidney staging, copper-hepatopathy risk, allergy exclusions. These are labeled separately from NRC values so you know what's reference and what's clinical policy.
The three specialized pipelines
The engine routes automatically based on the profile. Here's what each pipeline does differently.
Pediatric and orthopedic development
Puppies don't grow in neat age buckets. Their metabolic needs follow a continuous curve. Small errors in calcium or phosphorus during growth can produce irreversible orthopedic damage - the engine treats this window as higher-stakes than adult maintenance.
Protocol: exponential maturity model
Maintenance Amount = 130 × (Body Weight in kg)^0.75 // NRC 2006 Table 15-2
Growth Multiplier = 3.2 × (e^(-0.87p) - 0.1) // p = current / expected adult weight
DER = Maintenance Amount × Growth Multiplier // floored at maintenance (1.0×)
The growth curve scales the NRC maintenance amount (not resting energy): a young puppy needs roughly twice maintenance, easing toward maintenance as it approaches its expected adult weight - so targets don't swing abruptly at an arbitrary age cutoff. Model coefficients are an internal policy layer; NRC 2006 Table 15-2 remains the reference for nutrient targets.
Adult maintenance and genetic safeguards
A 20 kg Bulldog and a 20 kg Border Collie burn energy at different rates. Allometric scaling handles the baseline. Breed is applied as reviewable context - not as a substitute for measured intake.
Protocol: allometric scaling with modifiers
RER = 70 – (Body Weight in kg)^0.75
DER = RER – [Activity – Breed – Neuter – BCS – POMC]
Every modifier is transparent. You can see what's being applied and why. Two dogs with the same body weight can still end up with different targets - because the engine uses the full 8-point profile, not just weight.
Geriatric preservation and renal health
Senior dogs lose muscle mass. Unless they have confirmed kidney disease, they often need more high-quality protein - not less. The engine separates healthy aging from clinical renal failure rather than treating all senior dogs identically.
Protocol: sarcopenia defense
Target = Base Protein Requirement – 1.3 (Sarcopenia Boost Factor)
The boost applies for healthy seniors. When CKD staging is confirmed, the engine switches to IRIS-aligned guidance as a separate, labeled clinical policy - and makes the trade-off between protein and renal protection explicit, so you can review it with your veterinarian.
The safety gauntlet - before ingredients are touched
Once the baseline is set, the engine locks in constraints before any ingredient is evaluated. Hard ceilings come from NRC. Additional clinical policies, where applied, are labeled separately so you know what's reference and what's judgment.
Metabolic multipliers
Energy and context adjustments based on the dog's profile. Breed-specific considerations are labeled as separate assumptions - not NRC values - so you can see what's reference and what's interpretation.
Orthopedic ratio locks
Puppy growth is sensitive to calcium and phosphorus balance. The engine flags Ca:P ratio directionality and checks against life-stage-appropriate NRC growth targets rather than applying a single adult ratio across all life stages.
NRC Safe Upper Limits
Where NRC provides SUL values, the engine enforces them as hard ceilings in the same energy-based units as the targets. This matters most for selenium, vitamin D, copper, and iodine - minerals where the gap between adequate and toxic is narrow.
Renal and hepatic overrides
When a veterinarian has confirmed CKD staging, IRIS-aligned guidance applies as a separate, labeled clinical policy. These sit above NRC table values when active - and the engine marks them as such so your vet can review the logic.
Ingredient data quality flagging
If an ingredient is missing nutrient data for an essential target, the engine flags it rather than guessing. You'll know which ingredients are incomplete before the recipe is used - not after.
The Meal Planner - grams, not guesses
Steps 1 and 2 establish what your dog needs. Step 3 calculates how much of each ingredient delivers it. You pick the foods. The engine calculates the amounts - no spreadsheets, no unit conversions.
What goes in the bowl - and what each category does
Muscle meat
Core protein source. Provides essential amino acids, B vitamins, and zinc. Portions are scaled to meet your dog's daily protein requirement without exceeding the caloric target.
Raw meaty bone
Calcium and phosphorus delivery. The engine tracks the meat-to-bone ratio of each RMB option because the same weight of chicken necks and beef ribs delivers different minerals. Ca:P balance is maintained for the dog's life stage, not a universal adult target.
Liver
Concentrated source of vitamin A, iron, and copper. Because liver is nutrient-dense, the engine limits the portion to a range that meets targets without approaching hypervitaminosis A territory.
Secreting organ
Provides DHA, EPA, and fat-soluble vitamins that muscle meat alone can't supply. Portions are balanced against the day's total fat and energy targets.
Vegetables and fruit (model-dependent)
Added only when the feeding model includes them. If the model excludes produce, the engine doesn't force it into the ration.
The calculation
Portion (g) = Dog's DER – Category Ratio – Ingredient Nutrient Density?–
When you click Calculate, the engine looks up each ingredient's verified nutrient profile, applies your dog's feeding model ratios, and produces gram amounts that meet NRC targets within the safety ceilings. You get the result. Not the intermediate math. Not a percentage rule that might or might not apply.
What makes this different in practice
- Swap chicken breast for beef and click Calculate again - every portion adjusts to the new protein's nutrient profile.
- A 5 kg Chihuahua and a 35 kg Labrador choosing identical ingredients get very different amounts - because the engine uses each dog's profile, not a shared multiplier.
- If an ingredient is missing a nutrient value, the engine flags it rather than guessing. You'll see which gaps are real and which are data problems.
- Raw diet ingredients are high-moisture - a gram of liver is not a gram of dry matter. The engine applies moisture values before checking nutrient totals. Comparing wet-weight values to NRC dry-matter targets overstates deficiencies by up to 3– for high-moisture ingredients.
Bioavailability - what your dog actually absorbs
What's in the food is not what the dog uses. The engine applies absorption coefficients before checking nutrient totals.
Zinc (raw meat)
~35% AbsorbedCalcium (bone)
Varies by cut AbsorbedPhosphorus (bone)
Varies by cut AbsorbedIron (muscle meat)
Varies by form AbsorbedManganese (plant sources)
Varies by source AbsorbedWhere home-prepared diets go wrong - consistently
- Zinc is a common weak spot in home-prepared rations unless recipes are measured and checked against an energy-based reference (see Dillitzer et al., 2011).
- Calcium and phosphorus swing in opposite directions depending on how bone-heavy or bone-light the recipe is. Checking the ratio against a life-stage reference is safer than eyeballing.
- Manganese is low in most common raw ingredients by default. It doesn't show up unless you add a deliberate source.
Food first. Supplements second. Always.
When a nutrient gap is found, the engine doesn't immediately recommend a supplement. It checks whether a whole-food fix is practical first.
How the engine decides between food and supplementation
1. The Volumetric Ceiling
If fixing a zinc gap with whole food requires a physically impractical volume that would cause digestive distress, the engine switches to a measured supplement. Food first doesn't mean food-only-at-any-cost.
2. The Caloric Anchor
If the dog is already at their caloric limit, adding more food to close a gap causes weight gain. The engine anchors the calories and bridges the gap with a zero-calorie supplement recommendation instead.
3. Precision at the milligram level
For trace minerals like selenium and copper, whole-food adjustments are imprecise. When the gap is small and the safety margin is narrow, a measured supplement is the more controllable option. Outputs are estimates - the goal is repeatability, not false precision.
The clinical toolkit - what you get beyond the calculation
Real-time micronutrient audit
As you build a recipe, the Deficiency Meter checks each ingredient against NRC reference targets for 43 micronutrients. You see where the recipe is short on zinc, copper, or manganese before you finalize it - not after the dog has been eating that way for months.
Up to 4-week meal planning
Build one day, then clone it to any other day in one click - or clone an entire week in bulk. Four weeks of balanced meals in under 15 minutes. Every gram target carries your dog's profile, not a generic average.
Exportable vet report
Generate a consolidated shopping list and a printable NRC-backed nutrition report for your veterinarian. The report shows every nutrient, every target, and every gap - in units your vet can work with.
Symptom journal
Log observations - stool quality, energy, coat changes, itching. The engine maps each entry to that day's exact meal composition. You get a traceable correlation map, not a generic score. Exportable to your vet alongside the nutrition report.
Symptom correlation - connecting observations to ingredients
When you log an observation, the engine maps it to that day's macro ratios. If a loose stool flag lines up with a high-organ day or a new protein introduction, that connection surfaces. You don't have to guess.
| Severity tag | Observation | Engine action |
|---|---|---|
normal |
Ideal or firm stool, stable energy | Day marked stable. Macro ratios logged as baseline. |
info |
Minor softness, mild itching | Flag recorded. Overlay checks for elevated organ percentage that day. |
warning |
Unformed stool, repeated itching | Correlation check against bone percentage, meat percentage, new proteins introduced. |
danger |
Liquid or diarrhea, lethargy | Traceable trigger identified against that day's organ percentage and ingredient list. |
No arbitrary scoring. Just a date-based overlay.
Each symptom log is matched to that calendar day's macro breakdown. If a danger flag for liquid stool coincides with a high-organ day or the introduction of a new protein, the engine surfaces that as a traceable connection. You provide the observations. The engine provides the map.
Built for accuracy. Reviewed by a veterinarian.
This page explains the calculation assumptions and where the engine's boundaries are. It's a transparency layer - not a guarantee. For dogs with chronic disease, diagnosed conditions, or puppies of large and giant breeds, discuss recipe changes with your veterinarian alongside this output.
See the engine at work on your dog's recipe
Enter your dog's profile - results in under 60 seconds
Start freeRelated resources
- New to NRC 2006? Start with our plain-language explainer →
- Want the technical deep-dive? Explore the NRC standards guide →
- Veterinary professionals: Request our technical white paper → (Available upon verified request)
People also ask
How does Raw & Well use NRC 2006 in its calculations?
The engine uses NRC 2006 requirement classes (MR, AI, RA) and Safe Upper Limits, expressed per 1,000 kcal of metabolizable energy. Energy needs are calculated using allometric scaling (\(BW^{0.75}\)) rather than linear weight multipliers, then each nutrient is checked against NRC targets in those same energy-based units.
Is the scientific engine the same as a veterinary diet prescription?
No. Raw & Well checks recipes against NRC 2006 reference targets and flags gaps. It's a nutritional screening tool, not a medical device and not a prescription. For dogs with diagnosed conditions, the engine's output is a useful starting point for a veterinary conversation - not a replacement for one.
What is metabolic scaling and why does it matter?
Energy needs scale at \(BW^{0.75}\) - not linearly with weight. A 5 kg Chihuahua needs more calories per kilogram than a 40 kg Labrador. Percentage-based rules ignore this. The engine accounts for it, which is why its targets differ from calculators that apply a fixed percentage across all sizes.
Frequently asked questions
Is this aligned with NRC 2006?
The engine uses NRC 2006 requirement classes (MR, AI, RA) and Safe Upper Limits, expressed per energy intake (per 1,000 kcal ME). NRC doesn't approve individual recipes - we treat it as a reference for targets and ceilings, not a blanket guarantee.
Why are Raw & Well's numbers different from other calculators?
Most calculators use linear scaling: weight – multiplier. Raw & Well uses \(kg^{0.75}\) based on Kleiber's Law. Small dogs need more per kg; giant dogs need less. The difference is largest at the extremes, which is where simple rules are most likely to be wrong.
Can I trust this over my veterinarian's advice?
No - and it's not designed to replace your vet. It's designed to give your vet something concrete to work with. Share the output at your next appointment. Many veterinarians find precise nutrition records easier to work with than vague descriptions of what the dog eats.
What's the difference between this and the NRC explainer?
The NRC explainer answers "what is NRC 2006?" This page answers "how does Raw & Well use it?" They work together - start with the explainer, review the math here.
Is this appropriate for puppies and seniors?
Yes, with important caveats. The engine has separate logic for growth and senior phases. For large-breed puppies, calcium and phosphorus balance matters more than for adults. For seniors with confirmed kidney disease, IRIS staging applies as a labeled clinical policy. Both benefit from veterinary review alongside the engine's output.
Where can I see the full technical documentation?
The complete clinical specification - formulas, constants, breed multipliers, safety caps - is available in our technical white paper for veterinary professionals and technical users. Request it here.
See what the math finds in your dog's recipe
Enter your dog's profile and ingredients. The engine runs the calculation and shows you where the gaps are - with specific fixes, not generic advice.
Start your 7-day free trial