About WorkCapacity
Measure what matters
Fitness is supposed to be measurable.
That definition is precise. It is grounded in physics.
But most tools don't measure it directly.
CrossFit defines fitness as:
increased work capacity across broad time and modal domains
They measure time.
They measure reps.
They measure scores.
These are useful, but they are proxies.
They do not measure work.
They do not measure power.
They do not measure the thing the definition actually describes.
The Missing Layer
The most direct way to represent fitness is simple.
power on the y-axis
time on the x-axis
the area under the curve represents fitness
The higher and broader the curve, the fitter the athlete.
This is not new.
What's been missing is a system that makes it practical to compute.
Why Now
Agents change what matters.
Humans have never optimized directly for this metric.
Not because it isn't correct, but because it is abstract.
People respond to:
- Leaderboard positions
- Finish times
- Visible wins
AI does not.
Agents do not care about dopamine or simplicity.
They ruthlessly optimize toward defined objectives.
If you give them a clean signal, they will use it.
That shift changes what matters.
What WorkCapacity Is
A physics-based measurement layer.
WorkCapacity is a system for measuring and improving fitness using physics-based metrics.
At its core is Power Oracle.
Power Oracle is a lightweight API that:
- -> accepts workout structure, time, and athlete data
- -> computes total mechanical work
- -> computes average power output
- -> returns structured, machine-readable results
It is designed to be simple, precise, and usable by both humans and software systems.
The goal is not to create another tracking app.
The goal is to create a measurement layer.
Not Just A Number
Workouts are structured systems.
A single average power number is not enough.
Workouts are not uniform.
They have structure.
Intervals, transitions, fatigue, pacing - these all matter.
Power Oracle treats workouts as structured systems.
It measures:
- Per-segment work and power
- Consistency and dropoff
- How output is produced, not just the final result
This allows agents and coaches to reason about performance, not just record it.
Designed For Agents
Built for a new kind of user.
WorkCapacity is built for a new kind of user.
AI systems that:
- Plan training
- Evaluate performance
- Adjust decisions continuously
These systems need:
- Objective signals
- Consistent definitions
- Repeatable computation
Power Oracle provides that.
It exposes fitness in a form that can be queried, compared, and optimized.
The System
Measurement is the first layer.
Measurement is the first layer.
On top of it sits a larger system:
- an athlete model
- a coaching engine
- a feedback loop
- a programming system
Each layer builds on the one below it.
01
Measure
02
Model
03
Decide
04
Adapt
The Vision
A complete AI coaching platform.
WorkCapacity can evolve into a complete AI coaching platform.
A system that can:
- Model an athlete over time
- Prescribe training
- Track adaptation
- Optimize decisions across weeks, months, and seasons
The long-term goal is simple:
develop the fittest athletes in the world through systematic optimization
The milestone:
the first CrossFit Games athlete coached primarily by AI
In short
Measure what matters.
Make it computable.
Let systems ruthlessly optimize it.