Which reference values are used?

Ranges from the literature, with the caveat that most depend heavily on speed and stature. They sit beside the measurement as orientation, not as cut-offs, and nothing is labelled abnormal.

Can I supply my own norms?

Yes, in the local installation through a norms file. For a service working with a defined population that is more useful than ranges drawn from recreational runners.

Can I get at the raw data?

The measurement exports as JSON with the settings, the scaling used, the cycle count and the spread per measure. Without patient details, so a measurement can be shared or recomputed without a name travelling with it.

Which pose estimator is underneath?

RTMPose with the Halpe26 point set, through onnxruntime. Apache-2.0, so usable without licence restriction, and it runs in the browser. On top of the raw detection sits a correction that holds bone lengths constant across the clip; that removes fifteen to twenty-two per cent of the angular error around foot and ankle.

All running gait analysisFor practices

The limitations first, the price later

You assess whether a method supports the weight put on it. So the limits are at the top of this page and not behind a disclosure triangle.

What a single camera cannot do

Rotation about the long axis of the limb is not recoverable from one view. What is reported as hip adduction in the frontal plane is a projection, and an internally rotated hip changes that projection without changing the adduction. If rotation is the question, the answer is a multi-camera setup, and that is a different installation from a tripod beside a treadmill.

Foot and ankle are the weakest part. The model’s heel and toe points sit on the shoe, not on bone, and a thick midsole shifts the measured foot strike angle systematically. Rearfoot / midfoot / forefoot classification is therefore screening grade and wants calibrating against the footwear in front of you.

Absolute ground contact time from video underestimates by roughly five to ten per cent, because first contact and final toe-off fall between frames. The left-right difference and the change over time are unaffected: that error acts the same way on both sides.

How an event is established

Initial contact underpins nearly every temporal measure in the report. It is not determined by one method but by three, and only counts when at least two agree.

1
Foot height
The minimum of heel and toe against an estimated floor line. Robust, but blunt: near the minimum the foot barely moves, so the exact frame is hard to pick.
2
Vertical acceleration
The peak in the second derivative at landing. This is the sharpest of the three and supplies the timing; the other two only decide whether an event exists.
3
Foot relative to pelvis
The Zeni method: the extreme of the horizontal foot-to-pelvis distance. Works even when the floor is out of frame.

What has and has not been validated

Against a simulated runner with known ground truth: initial contact within zero frames, toe-off within about two, cadence within one step per minute, vertical oscillation within one per cent, and zero asymmetry on a symmetric runner. That tests the arithmetic end to end, not the world.

What does not exist is a study on real footage against force plates or an optoelectronic system. It ought to, and the protocol is written: concurrent measurement against a reference to establish systematic bias per measure; within-session and between-rater repeatability; and from those, the smallest change that reliably clears the noise. Until that exists, this page states what there is and nothing beyond it.

Why this is not a medical device

The software measures and assesses the quality of that measurement. It does not link values to a condition, predict anything, or propose treatment. That places it outside the definition of medical device software, and Rule 11 does not arise.

The reasoning is written out against the five decision steps of MDCG 2019-11 and is a public document. It also states what would change with an interpretation module included: Class IIa, with the route that goes with it. That module was built and deliberately removed again.

Frequently asked

Which reference values are used?: Ranges from the literature, with the caveat that most depend heavily on speed and stature. They sit beside the measurement as orientation, not as cut-offs, and nothing is labelled abnormal.
Can I supply my own norms?: Yes, in the local installation through a norms file. For a service working with a defined population that is more useful than ranges drawn from recreational runners.
Can I get at the raw data?: The measurement exports as JSON with the settings, the scaling used, the cycle count and the spread per measure. Without patient details, so a measurement can be shared or recomputed without a name travelling with it.
Which pose estimator is underneath?: RTMPose with the Halpe26 point set, through onnxruntime. Apache-2.0, so usable without licence restriction, and it runs in the browser. On top of the raw detection sits a correction that holds bone lengths constant across the clip; that removes fifteen to twenty-two per cent of the angular error around foot and ankle.

The evidence page has the rest

How it measures, what has been tested, what has not been tested yet, and the sources underneath.

Read the evidence

Read further

Running gait analysis for physiotherapists
Measure the running pattern from a slow-motion clip, put it in the notes, and at review see whether anything actually changed. From € 20 an analysis.
For podiatrists: what video gait analysis does not measure
Honest answer: foot and ankle are the weakest part on 2D video. What this adds is what happens above the foot, and what changes when the orthosis goes in.
Running gait analysis for coaches
Cadence, overstride and contact time as numbers, so that across a season you can show what changed instead of asserting it. From € 20 an analysis.
For running stores: not the gait analysis you already do
This does not tell you which shoe fits. It does show what measurably changes in the running pattern between two shoes: which is a service you can sell on its own.
Pricing for running gait analysis
Unlimited for € 129 a month, or from € 20 an analysis on a bundle. Bundles do not expire. First analysis free.
Evidence: how it measures and what has been tested
Which model sits underneath, how gait events are established, what has and has not been validated, and why this is not a medical device.
Recording protocol for running gait analysis
Camera position, frame rate, clothing and duration. The recording determines the measurement more than the software does, and this is what has to be right.
Compared: camera, IMU sensors, a motion lab or the naked eye
What camera-based analysis can and cannot do against IMU sensors, a gait lab and assessment by eye. Including the places where the others win.
Try it: the first analysis is free
One complete analysis with the report, no watermark and no card details. You need a twenty-second slow-motion clip and nothing else.
Bike fitting with the camera you already own
Measure your client on the trainer from an ordinary clip. Joint angles per crank revolution, adjustments in millimetres, a report for the file. From € 28 an analysis.
Overpronation: what it is and what it does not predict
What pronation is, why sorting runners into neutral, over- and underpronating says less than long assumed, and what you can usefully look at instead.