Features

Engineered from
first principles.

Every decision VirtueFit makes traces back to your body. Not a template, not an average, not a guess. A closed-loop system that reads, plans, adjusts, and learns.

0data points per scan
0variables per programme
0ms response time
0hr adaptation cycle
01 Computer Vision

Body Scanning

A single phone camera captures a full 3D reconstruction of your physique. Pose estimation, depth inference, and mesh generation run in sequence to produce a body model accurate enough to derive composition, circumference, and postural data, all processed on-device in under sixty seconds.

  • InputRGB video stream, 30fps, 8-12 second capture window
  • PipelinePose landmarks → depth map → SMPL mesh → measurement extraction
  • Output3D model, body fat estimate, lean mass, 14 circumference measurements
  • RefreshDelta-comparison across scans for visible change over time
scan_pipeline.process()
Pose detection
96%
Depth inference
91%
Mesh generation
88%
Composition model
94%
02 Inference Engine

Adaptive Programming

Your training programme is not written once. It is generated session by session from a constraint-satisfaction model that balances progressive overload, fatigue accumulation, movement pattern distribution, and your personal recovery profile. Every completed workout feeds back into the next generation cycle.

  • ModelMulti-objective optimisation across volume, intensity, frequency
  • InputsSession logs, RPE signals, sleep patterns, scan deltas
  • SplitsPush/pull/legs, upper/lower, full-body, or fully custom topologies
  • CadenceRe-generation triggered post-session or on schedule conflict
engine.generate_programme()
User Profile
Goals Capacity History
Volume Intensity Frequency Recovery
Programme
03 Feedback Loop

Continuous Learning

Every interaction is a signal. Completed sets, skipped exercises, logged meals, scan results, and direct questions all feed into a unified profile that sharpens with use. The system does not assume steady state. It models you as a moving target and adjusts its predictions accordingly.

  • SignalsSession completion, RPE, adherence rate, scan deltas, chat queries
  • CycleIngest → weight → re-rank → regenerate within 24 hours
  • DriftDetects plateau, over-reaching, and preference shifts automatically
  • ContextTime, equipment, energy, schedule — all factored before generation
feedback.ingest_cycle()
Scan Train Log Learn Adjust
AI
04 Unified Model

Nutrition Intelligence

Calorie and macronutrient targets are not static numbers. They are derived from your latest scan, your current training block, your logged meals, and your stated goal. The system rebalances daily, accounting for training load, recovery demand, and adherence patterns so your fuel matches your output.

  • DerivationBMR from scan → TDEE from activity → surplus/deficit from goal
  • MacrosProtein, carbohydrate, fat targets rebalanced per training day type
  • MealsAI-generated suggestions matched to preference, time, and target
  • LoggingNatural language or barcode, parsed to macros in under two seconds
nutrition.derive_targets()
Protein2.2 g/kg
Carbohydrate4.8 g/kg
Fat0.9 g/kg
Daily target2,840 kcal
05 Natural Language

Conversational Coach

Ask anything. The coaching layer has full read access to your profile, your scan history, your programme, your nutrition log, and your check-in data. Responses are grounded in your numbers, not generic advice. Swap a session, query a macro split, ask why a weight was prescribed. It answers from your context.

  • ContextFull profile, scan history, programme state, nutrition log
  • ActionsSwap workout, adjust meal, explain prescription, surface insight
  • LatencyFirst token in ~120ms, full response streamed in under two seconds
  • ToneDirect, specific, evidence-cited — not motivational filler
coach.respond(query)
Why did my squat volume drop this week?
Your scan on Monday showed a 1.2kg lean mass increase. Combined with elevated RPE on Thursday's session (8.5 avg vs 7.2 norm), the system reduced squat volume by 15% and redistributed to accessories to manage cumulative fatigue while preserving the stimulus.

The full system. One app.

Scanning, programming, nutrition, and coaching, unified under one engine that improves every time you use it.

Download on the App Store