How BikeFit-IA Uses Artificial Intelligence to Analyze Your Position

BikeFit-IA isn't "just another AI" that detects your joints and shows you angles. Behind every analysis, there's real diagnostic work — and that's what makes the difference.

What most apps do

The majority of bike fitting apps follow the same pattern:

1. You film or photograph your position
2. An algorithm detects your joints
3. The app displays your angles: "Knee: 145°"

And that's it. You walk away with numbers, not knowing whether they're good or bad, or what to do about it.

Pose detection models are widely accessible today. Displaying an angle — any app can do that. The real question is what happens next.

What BikeFit-IA does differently

A diagnostic engine, not just an angle display

When BikeFit-IA analyzes your video, pose detection is only the first step. Your angles then feed into a diagnostic engine built specifically for bike fitting:

• Your 5 joints are cross-referenced. If your hip is too closed AND your back is too, the system identifies it as a single problem (position too compact), not two separate issues. One adjustment can fix everything.
• Thresholds are tailored to your discipline. A triathlete in aero position and a weekend road cyclist don't share the same optimal angles. BikeFit-IA knows the specific targets for road, TT, triathlon, gravel, MTB, and track.
• Recommendations are prioritized. You can't fix everything at once. The system identifies which adjustment will have the greatest impact and tells you that first.

Thresholds grounded in science

Our reference ranges are based on published research in cycling biomechanics. Peer-reviewed science, not opinions or guesswork.

When the system tells you an angle is out of range, it's because research has established that this range minimizes injury risk and optimizes pedaling efficiency for your type of riding.

Pedaling style detection

When you pedal, your ankle doesn't stay still. Some riders pedal on their toes (toe-dippers), others let the heel drop (heel-droppers), and others have a fluid range of motion (anklers).

This detail changes how all your angles should be interpreted. A toe-dipper with a knee at 148° doesn't have the same biomechanical reality as an ankler with the same angle. BikeFit-IA detects this automatically and factors it in. No competitor does this.

Temporal distribution

Other apps show "min: 142° / max: 158°". Anxiety-inducing. You see the extremes and panic.

BikeFit-IA shows you how much time you spend in each angular zone. "92% of the time between 145° and 152° (optimal). The 158° peak lasts 0.04 seconds per pedal stroke." Context. Nuance. You immediately understand that your position is fine.

An AI Coach that knows YOUR data

BikeFit-IA's AI Coach isn't a generic chatbot. When you ask it a question, it has in front of it:

• Your 5 joint angles measured dynamically
• The diagnostics generated by the engine (what's optimal, what's not)
• The correlations identified between your joints
• Your discipline and specific targets

When you ask "Why does my back hurt?", it doesn't recite a Wikipedia article. It looks at your data and tells you: "Your torso is at 35° while the range for your discipline is 40°–55°. Combined with your closed hip, this indicates a position that's too compact, overloading your lower back. The priority adjustment is to raise your handlebars."

That's the difference between reading a book about bike fitting and having an expert look at your position.

Why this approach is different

What we don't claim to do

We're transparent about our limits:

• We don't replace a professional bike fitter for complex cases — chronic injuries, significant asymmetries, known conditions.
• Analysis quality depends on the video. That's why we developed a calibration system (AprilTag) that verifies your camera alignment and automatically detects parallax errors.
• We don't claim to solve everything in one video. Bike position is refined through iterations — and that's exactly why we allow repeated analyses: adjust, re-film, compare.

In a nutshell

BikeFit-IA is:

• A diagnostic engine that cross-references your 5 joints and identifies the root cause of your position issues
• Science-based thresholds tailored to each discipline, sourced from published research
• Pedaling style detection for accurate interpretation of your angles
• Temporal distribution that replaces anxiety-inducing min/max values with real context
• An AI Coach that knows your data and gives you concrete guidance

Not just angles. A diagnosis.

---

BikeFit-IA — Artificial intelligence serving the cyclist, not the other way around.