Pearl One hello@pearlone.app
4:21:39 PM the collar flags it
Ninety-five seconds from Pearl's collar on October 2, 2026. The flat stretch is Pearl at rest. Everything after it is a seizure.

Pearl's seizures don't wait for someone to be watching.

Pearl One is a prototype collar I built after my cat, Pearl, started having seizures. It records how she moves and sends an alert to my phone when that movement looks like one of her past seizures. There's no camera involved. It has been tested on one cat so far: Pearl.

About Pearl

Pearl has seizures. They arrive without warning, at any hour, and often in clusters. For a long time the only record I had was my own memory of what I'd seen, and the seizures I didn't see left no record at all.

Today she's on medication and doing well. Her vet manages her treatment; the collar hasn't been used to make medical decisions. She has gone almost two weeks without a seizure as I write this, but because they come in clusters, I'm still watching.

What she wears

A camera only sees what's in frame. A collar goes where the cat goes. Pearl One is a small motion sensor in a 3D-printed case on an ordinary collar. When Pearl's movement looks like one of her past seizures, it sends an alert to my phone.

It also keeps an honest record. If any part of a recording is missing, the record says so instead of quietly papering over it.

October 2, 4:21 in the afternoon

This is the clearest case I have. I was in the room, the home camera was recording, and Pearl was wearing the collar.

Real video of a seizure

The clip shows Pearl having a seizure on our home camera. It runs about a minute and a half with no sound, and some people will find it upsetting. The timeline beside it tells the same story without the video.

  1. On the home camera, Pearl's movement changes.
  2. The collar flags the motion. The alert it sends names this same second.
  3. I tap "seizure" in the app.
  4. I mark the end of the seizure.

Four separate records agree on 4:21:39: the camera's clock, a second recording of that clock, the collar's own data, and the alert. Nothing the collar recorded in that session was lost.

See the full case, with the motion data and every timestamp

The record so far

8seizures captured on the collar between September 15 and October 13, 2026, each confirmed by me
7of those eight have a record of me reporting that the alert worked
1false alert confirmed: Pearl scratching, on October 7
473+hours of motion recorded across 80 sessions on 32 days

Every session reports how complete it is. Gaps, recoveries and incomplete sessions stay in the record instead of being cleaned out, and early failures are kept because they shaped how the system works now.

What this doesn't show yet

  • It's one cat. Nothing here says how it would work for another.
  • Pearl One isn't a medical device. It doesn't diagnose, treat, prevent or predict anything.
  • The numbers above are counts from one household, not accuracy rates.
  • The hardware is a prototype. Comfort, battery life and safety for everyday use still need proper testing.
  • No vet or researcher has reviewed this work yet. That's the next step.

What comes next

The next step is a pilot with a second cat in another home, run without a camera, with the rules set before it starts and an independent review at the end. I'm looking for:

  • veterinary and research partners who work with feline seizures
  • funding for that pilot and a sturdier version of the collar
  • people who have taken a small hardware product from prototype to shelf

Who's building this

I'm Tim Pearce, Pearl's person. I've spent fifteen years building technology systems in hospitality, most recently as an enterprise architect. I built Pearl One in my own time because I wanted a record of what was happening to her when I wasn't looking.

Get in touch

If you work with cats, seizures, research or small hardware and this interests you, I'd like to hear from you.

hello@pearlone.app