What 250+ Phone Camera Body Scans Revealed About How We Move

Can a phone camera detect movement patterns that people don't know they have? We built a browser-based tool to find out, and ran it on more than 250 people.

The Setup

Four movement tests (shoulder raise, toe touch, overhead squat, leg raise), each designed to reveal a different pattern. A phone propped up at arm's length, AI tracking your body in real time, with the camera analysis running entirely on your device: your video never leaves your phone. A few minutes, no app download.

What the Data Shows

About this dataset: more than 250 assessments collected in 2025 with earlier versions of Previa, aggregated and anonymized. An early internal evaluation, not a peer-reviewed study. The current version has since logged 600+ individual measurements on top of this.

Average score: 61 out of 100. Out of 250+ people, only 12 had no findings at all; 95% showed at least one measurable pattern. Most land in the Moderate range.

The most common findings, ranked by prevalence:

64% showed forward head posture: the neck drifting forward instead of staying stacked over the spine. If you work at a desk, there's a good chance you have it and don't know. The overhead squat turned out to be the most revealing test: three out of four people couldn't do it without compensating. Most thought their form was fine.

Your Body Is Probably Cheating

One user raised his leg to 80 degrees. That looks good. But when we looked at the full picture, his torso was leaning 27 degrees to the side to make it happen.

His actual hip mobility: 53 degrees. His body had been working around a significant restriction, probably for years. He had no idea.

This wasn't an outlier. 73% of the assessments showed the same thing: the body compensating to reach a range it can't actually achieve cleanly. What looks like "good mobility" is often the body finding a workaround.

Think of it like Oura's "time in bed" vs. "actual sleep." The raw number looks fine. The compensation-corrected number tells a different story.

You're Probably Moving Older Than You Are

From the overall score and compensation data, we calculate a "Movement Age": how old your body moves compared to your actual age. In this early evaluation, 7 in 10 people moved older than their actual age, by 7 years on average. A 28-year-old moving like a 35-year-old.

What Surprised Us

We expected flexibility to be the main issue. The data pointed somewhere else: compensation. People can move far enough. Their bodies just cheat to get there.

One user with 30 years of chronic back pain ran the assessment. We flagged exactly what his physiotherapist had independently found: his right hip. He wrote us afterward: "First time something digital actually matched what my physio told me."

Another user took his results to his osteopath. The osteopath confirmed the same patterns, the asymmetry, the compensation side, the restricted hip, without having seen our report. That was the moment we knew we were onto something.