Our Methodology
Coaching judgment. Structured rules. Transparent AI.
KarvedUp is designed for generally healthy adults age 18 and older pursuing strength, physique, or fat-loss goals. Our method prioritizes a plan you can consistently execute, measures what happens, and makes small adjustments from sustained trends.
Progress earned
Training progresses from logged performance—not novelty for its own sake.
Consistency first
The plan should fit the athlete's real schedule, equipment, and ability to follow it.
Rules before rhetoric
Core targets, trends, portions, and progression suggestions use structured calculations.
AI with boundaries
AI helps personalize and communicate the plan; it does not turn estimates into certainty.
Progress without rebuilding the plan every week.
Programming blends AI-assisted plan design with deterministic progression and periodization rules.
Start with the athlete's reality
Goals, training experience, schedule, available equipment, and self-reported movement restrictions shape the split and exercise selection. AI assists with the initial structure and exercise choices inside those constraints.
See this in Your Plan → Training.
Use a repeatable training block
Programs run in four-to-six-week mesocycles. Training effort follows a programmed repetitions-in-reserve wave, with a lower-volume deload in the final week. Reusing session templates makes performance comparable from week to week.
Progress from completed reps
Load suggestions use double progression: reach the top of the prescribed rep range, and the next suggestion adds five pounds and returns to the bottom of the range. For loads below 30 pounds, the system adds reps first, before making a five-pound jump.
Example
| Week | Weight | Reps | Why |
|---|---|---|---|
| Week 1 | 185 lb | 10 reps | Top of the 8–10 rep range |
| Week 2 | 190 lb | 8 reps | +5 lb, back to the bottom of the range |
Illustrative numbers — your actual weights and rep ranges come from your logged sets.
See this in Workout, on your next prescribed set.
Review the week, then progress the plan
The weekly review reads every logged lift and proposes one change per lift. When every set reached the top of the rep range at the planned effort, the weight goes up by a small step sized to the athlete's experience. Otherwise the weight holds and the target is more reps. Poor recovery from check-ins, or reps dropping two weeks in a row, cut a set or lighten the lift, and lifts flagged as painful get lighter. A ready-to-push week with good recovery can add a set within volume limits for the athlete's level. The training changes apply only if the athlete chooses to apply them, separately from nutrition.
See this in Weekly Review.
Adjust the smallest useful variable
Exercise swaps and same-day adaptations are available when equipment, time, or circumstances change. Readiness check-ins inform coaching context; they do not silently rewrite today's workout prescription.
See this in Workout, using Swap or Adapt Workout.
Estimates become useful through trends.
No calorie equation can directly measure an individual's metabolism.
KarvedUp starts with an established estimate, then evaluates real logged outcomes.
Baseline targets
Resting energy starts with the Mifflin–St Jeor equation, then applies an activity estimate and a goal-specific calorie adjustment. Protein and fat targets follow consistent product rules; carbohydrate fills the remaining energy.
See this in Your Plan → Nutrition.
Home-meal portions
AI proposes foods and recipes. The daily meal plan reads nutrient densities from stored FatSecret food ids, then solves gram portions in code using 4/4/9 calorie math and consistency checks. Food search, scans, swaps, and restaurant items use FatSecret. A barcode uses Open Food Facts only when FatSecret has no hit.
See this in Nutrition, on your meal cards.
Weekly trends
The review engine uses a seven-day rolling weight trend. It requires at least four weigh-ins and sufficient nutrition logging before proposing a calorie or macro change; otherwise, current targets are held.
See this in Weekly Review.
Approval
Weekly calorie and macro changes are presented before they take effect. The athlete answers nutrition and training separately, and each defaults to keeping the current plan, so either can be applied without the other. Adjustments are intentionally incremental rather than reactions to a single weigh-in.
See this in Weekly Review.
Where estimates remain: Restaurant meals, natural-language food logs, and foods without a reliable database match may use AI-assisted estimates. Rebalancing applies only to meals that remain unlogged and available to adjust. These values are planning estimates, not laboratory measurements.
What AI does—and what code controls.
AI proposes options and explanations; rule-based code calculates and enforces the numbers.
AI-assisted
- Plan structure and exercise selection
- Meal and ingredient ideas
- Workout adaptations and exercise swaps
- Natural-language coaching and explanations
- Parsing natural-language food logs
Rule-based
- Baseline calorie and macro calculations
- Weight-trend and logging sufficiency checks
- Weekly macro adjustment rules
- Home-meal portion solving and rebalancing
- Mesocycle timing and load suggestions
Safety and scope
KarvedUp uses self-reported injuries, movement restrictions, allergies, and preferences as generation constraints. These are software guardrails—not clinical screening.
Do not push through pain. Stop the provoking movement and consult a qualified healthcare professional when pain is acute, severe, persistent, or otherwise concerning.
What we do not claim
- KarvedUp is not a live human coach or medical provider.
- Recommendations are not diagnosis or treatment.
- Food, calorie, and progress estimates are not exact measurements.
- AI output is not guaranteed to be error-free.
- Individual outcomes are not guaranteed.
Evidence informs the system. It does not replace judgment.
These sources inform broad principles and calculations.
KarvedUp's specific implementation and product rules have not been independently clinically validated.
American College of Sports Medicine, 2026
Resistance training prescription for healthy adults — position stand
DOI: 10.1249/mss.0000000000003897
American College of Sports Medicine, 2009
Progression models in resistance training for healthy adults — position stand
Mifflin et al., 1990
A predictive equation for resting energy expenditure — American Journal of Clinical Nutrition
International Society of Sports Nutrition, 2017
Diets and body composition — position stand
FatSecret Platform
FatSecret Platform API
Questions about how KarvedUp works?
We welcome questions, corrections, and thoughtful scrutiny of the methodology.
KarvedUp provides AI-driven training, nutritional, and caloric estimates for educational and informational purposes only. It does not provide medical advice, physical therapy, or clinical diagnosis. Always consult a qualified healthcare professional before starting any new diet or exercise regimen.