Schema is a language for describing training load. SchemaFit is the brand of tools built around it, beginning with SchemaFit Timer: describe, execute, model and study a workout as one structured object.
A workout is more than a list of intervals. It is a system that can be represented, transformed and understood.
What is the Schema language? Schema is a compact, human-readable language for describing physical training load: work and recovery, intensity, repetitions, tempo and how they change over time. A Schema can be executed by a timer, modeled, and compared with measured data.
30/210work and recovery duration@9.5target effort on a 0–10 scalet30movement tempo*5 #repeat five times, then finishA Schema can do more than repeat the same interval. It can shape a finish, draw a tempo wave, generate a repeatable challenge or leave room for what happens in the real session.
[base=18 race=22 kick=26]
120/0@t[base->race]
360/0@t22
120/0@t[race->kick]#Two minutes to build pace, six to hold it, then a controlled finishing acceleration. The shape of the effort is part of the prescription.
[lo=24 hi=36]
(60/0@t[lo->hi]
60/0@t[hi->lo])*5#Five continuous waves. Watch the heart-rate response lag behind each rise and settle on each descent.
random(
20/10@7 12/18@9
30/20@6
turb:8 dur:900 min:10 seed:31
)#Variable bursts for a sailing session, with a seed that lets you run the same virtual course again next month.
dur:600(~/0@r~*~)#Set a ten-minute window, then record the work and repetitions you actually complete. The result becomes a concrete session you can compare.
[dw=2 dr=2 N=8]
(dw/dr@r1
->[to:auto/auto@rN,delta:+dw/+dr@r+1]
->[to:auto/auto@r1,delta:-dw/-dr@r-1])*1`#Repetitions rise from one to eight, then descend to one. Work and recovery grow with the target, so each rung gets its own timing.
The same Schema connects the plan, the live session and the physiological response. That continuity is the foundation of the whole system.
Write the intended load once, including change over time and the variables that matter.
Materialize the plan for a person and discipline, then guide the session in real time.
Compare the model with the measured response and refine both the workout and the model.
SchemaFit Timer is where the language becomes a session. It combines precise timing, workout structure, heart-rate modeling, live telemetry and post-session comparison in a single offline-first instrument.
Run a Schema with real sensors, compare the measured response with your individual model, and refine that model from what actually happened.
Heart rate, zones, timing and spoken feedback during the session.
Practice Load, Rest, Tau ON/OFF and drift against temporary training curves. Improve the fit, reach 95%+, earn trophies and learn how each parameter changes the response — without touching real HR data or saved curves.
These are real screens from SchemaFit Timer: one system for authoring a Schema, running it, recording the response and testing the individual model against measured heart rate.
A two-minute, 50-repetition Schema was executed with live heart-rate telemetry. The result was then compared with the personalised response model inside the app.
A Smart Curve is more than an HR file. It is a saved training attempt with a persistent curveId, its Schema and analysis context. Link related attempts into a family and see how the athlete's response changes across passes.
Each curve has its own identity. A parentCurveId links a follow-up to its parent, so filenames and folder paths never define the history.
Family comparison keeps context. The root becomes Base; later attempts become Results. A family archive includes source CSV/FIT files and family.json. Related video is matched by curveId and shared separately.
The camera is part of the instrument, not an afterthought. It records the athlete and places selected Timer blocks directly into the video while keeping files small enough for everyday use.
Any Timer information block can become part of the recording: time, repetitions, heart rate, deltas, progress and other live telemetry.
01During the first five seconds after the start, face-aware metering inside the central 60% area establishes focus and exposure, then locks both. At that point the athlete is guaranteed to be in position.02Every three minutes the camera briefly corrects for changing light and locks exposure again — useful outdoors as evening light changes.03Manual exposure compensation remains available across a ±2 EV range, alongside a broad choice of recording formats.04The compact recording mode is engineered for very small files: a ten-minute training video can remain under 100 MB.SchemaFit is designed as a family of tools built around one shared representation of load — across disciplines, devices and research questions.
Specialized experiences can materialize the same abstract Schema through pace, power, cadence, terrain or movement tempo.
Predict a response, compare alternatives and search for a workout that fits a target, a time limit or an athlete's constraints.
Use Schema as an intermediate representation to translate comparable load between running, cycling, kettlebells and other domains.
Estimate individual response parameters from repeated sessions and make the next prediction more representative.
Reconstruct an explicit Schema from a coach's description, an existing protocol, telemetry or video. AI performs the interpretation; Schema provides an inspectable, editable and executable result that can be modeled, compared and reproduced.
The applications may look different because their domains are different. Underneath, they can exchange the same structured description instead of reducing every workout to a generic interval list.
SchemaFit did not begin as a speculative software idea. It grew from competitive sport, coaching, repeated self-experimentation and a community that has been recording comparable efforts over time.
Timers can count seconds. Sensors can record what happened. Training plans can describe intent in prose. But these three layers rarely share one precise, executable representation.
Schema was created to preserve the structure between intention and response: what changed, when it changed, how it should be materialised in a sport and what can be compared afterwards.
01Competitive practice exposed the limits of generic interval lists.02Community challenges made repeated, comparable efforts visible over months and years.03Modeling required a formal description of the input — not heart-rate data alone.04The language, Timer and analysis tools emerged as parts of one system.Nearly 3,000 published challenge attempts document real training histories across exercises, ages and ability levels. The collection includes rules, rankings, dates, repeated attempts and individual progress — a practical foundation for longitudinal analysis.
This is observational field data, not a controlled clinical trial. Its value is different: it reveals trajectories, generates hypotheses and shows what sustained practice looks like outside a laboratory.
Physiological measurements become more useful when the intervention that produced them is explicit. SchemaFit connects a computable protocol with execution, measurement and model revision.
01Describe protocol02Predict response03Execute session04Measure response05Compare & refineA Schema records timing, intensity, tempo, progression and variability in a machine-readable form that can be shared and rerun.
Repeated sessions support N-of-1 comparisons: the same athlete, controlled changes, explicit assumptions and inspectable outcomes.
Predicted and measured curves can be compared using correlation, timing offset and error measures instead of visual resemblance alone.
R/01How deliberate variation in tempo, intensity and interval structure affects adaptation and performance.
R/02Estimating time constants, drift and recovery from repeated sessions instead of population formulas alone.
R/03Comparing protocols by achieved stimulus, time cost and practical constraints.
R/04Testing when two very different activities can produce meaningfully comparable load profiles.
R/05Turning a target physiological trajectory into an executable workout for a chosen domain.
These are research directions, not settled claims. Community records are observational; heart-rate models are approximations; and correlation is not validation by itself. SchemaFit's role is to expose assumptions and make protocols structured, shareable and precisely repeatable — so stronger studies can be designed.
Practical manuals, language tutorials and research-oriented notes. English editions are paired with the original Russian documents.
The complete practical guide to the Graph, Schema editor, Smart Curves and attempt families, multi-Schema programmes, camera HUD, sensors, analysis and the Performance Optimiser.
A gradual, example-led introduction: from 30/30# to calibrated load, materialisation and controlled variability.
A structured tutorial covering atoms, groups, repeats, scopes, deferred values, progressions, ramps and random Schemas.
How the graph, Schema, heart-rate model, real execution and AI-assisted interpretation form one training cycle.
Tempo ramps, progressions, seeded randomness, sport-specific patterns and repeatable heart-rate comparisons.
The product is shaped by the same values as the language: explicit structure, useful freedom and no hidden noise.
Core training and analysis should not depend on a network connection.
The free version is a real scientific instrument, not a funnel filled with interruptions.
A model is presented as a model, with parameters that can be inspected, practiced, compared with reality and refined.