TODAY
Train with what you have.
Digital Coach builds the next session from your plan, recovery, personal response model, weekly dose, uncertainty, and the equipment available today.
Today's setup
Defaults for the next session.
ADAPTIVE CYCLE
Training phase
Complete training sessions to establish a progression baseline.
Readiness
Quick check-in before training.
MUSCLE ATLAS · 25 DIMENSIONS
Dimensional recovery map
Front and back recovery are shown as a clean muscle atlas. Every colored region is an independent model for readiness, weekly effective dose, 48h load and confidence, so a large group such as Chest, Back or Legs is never treated as one single recovery state.
TRAIN
Build today's workout.
Your saved plan supplies the constraints. Digital Coach optimizes the exercise portfolio and today's prescription inside them.
Today's session
Equipment mode and deload status can be changed here without altering your saved plan.
Saved workouts
Load an exact template or refresh its prescriptions from your latest performance.
LIBRARY
Exercise library
Search strength, warm-up, cardio, and cool-down movements by muscle or equipment.
SESSION MOVEMENT DATABASE
Warm-up, cardio, and cool-down
48 timed records classified by role, session focus, impact, age suitability, and equipment. Every record includes a local autoplay GIF and complete instructions.
PROGRESS
Training progress
Performance, effective dose, and the Digital Coach model learning from your own training.
Training volume trend
Total completed working-set volume across the last eight calendar weeks.
Muscle-group distribution
Fractional effective working sets by training region during the last four weeks.
Weekly effective-set targets
Direct work counts as 1.0 set and meaningful secondary work as 0.5, so overlap from compound lifts is not ignored or double-counted as full direct work.
Weekly coaching review
Stored decisions from each calendar week, including adherence, reported reps left, volume changes, and conditional deload signals.
Digital Coach model
Your N-of-1 model learns which exercises, weekly doses, and fatigue levels produce the best repeatable performance for you.
Strength records
Estimated performance adjusts an Epley-style estimate using the final-set clean reps-left report when available.
Workout history
Finished sessions and completed exercises.
MY PLAN
Plan and settings
Set the stable parts once. Day-specific choices stay on the Train screen.
Profile
Inputs that affect starting guardrails and workout generation.
Workout defaults
Used automatically unless you override the next workout on the Train screen.
Digital Coach
Controls how aggressively StrengthLab learns from your own N-of-1 training history.
Exercise preferences
Bias, reduce, or exclude exercises without changing the library.
Library and storage detailsDatabase counts, local media, prescriptions, and recorded sets.View
Programming controls
Edit sets, reps, load, and rest from any exercise detail.
Body composition
Manual local entries, because an offline web package cannot read Apple Health or Health Connect.
Backup and transfer
One-click export includes profile, workouts, logs, prescriptions, preferences, the Digital Coach model, custom exercises, and body measurements.
Firebase account
Private username/password login with user-scoped cloud backup. Local storage is isolated by the signed-in account.
Checking your StrengthLab account and cloud state.
Capability auditImplementation status for core app capabilities and V7 Digital Coach systems.View
Scientific basis and algorithm limitsEvidence priors, N-of-1 learning, and safety boundaries.Read
Population evidence becomes a prior, not a permanent prescription. V7 starts from evidence-informed dose, effort, rest, exercise-selection, and autoregulation guardrails, then gradually learns from repeated same-exercise and same-muscle observations from this account.
Digital Twin. Each exercise has a probabilistic response and fatigue model. Early predictions remain uncertain and close to conservative population priors. Repeated exposures reduce uncertainty. The model learns performance response, completion, fatigue cost, and which weekly effective-set doses have produced the best outcomes for each training region.
N-of-1 exploration. When two safe exercises are close in predicted utility, V7 may give a small bonus to the less-certain option. This is bounded exploration, not random exercise selection. Equipment, movement-slot, age, recovery, explicit exclusions, and safety constraints remain hard filters.
Set-by-set autoregulation. If an early working set materially exceeds or misses its rep target, or same-load reps collapse across sets, the next set can be adjusted immediately. The user can manually edit any set; user-edited values are not overwritten by the live coach.
Exercise Genome. Each movement is represented by pattern, primary and secondary muscle contribution, resistance family, support/stability demand, laterality, skill, setup cost, heuristic fatigue cost, joint-demand map, progression modes, and a modest long-muscle-length prior. These descriptors allow replacements and workout portfolios to be compared by function instead of name alone.
Constrained optimizer. A workout is selected as a portfolio under movement coverage, time, recovery, fatigue, equipment, weekly dose, continuity, preference, and uncertainty constraints. The model optimizes expected useful training rather than maximizing exercise count.
Exercise compiler. V7 does not fabricate new biomechanics. When a known movement is clearly too easy, it may compile a tempo, pause, or manual-resistance modifier that preserves the underlying movement and existing demonstration. Novel mechanics still require an explicit catalog exercise and appropriate media.
Digital Coach response/fatigue values are decision-model signals, not medical measurements. Evidence priors are based on contemporary resistance-training literature and are intentionally updated conservatively by personal data.