What ColorSense is

The device that knows
your wardrobe — hat to shoe.

ColorSense is a spectral sensor and AI wardrobe system that learns every piece of clothing you own and tells you what to wear, out loud, for wherever you’re going. Not “here are some options.” An actual answer. No screen required.

Point it at something. Say where you’re headed. It tells you what works and why — the way a stylist would, except it never sleeps, never judges, and has absolutely no opinion about how long you’ve owned that jacket.

Works on
👔 Shirts
🥥 Jackets
👖 Trousers
👗 Dresses
👟 Shoes
🎩 Hats
Four ways to teach it your wardrobe
Walk-through scan
Pan your phone slowly across your wardrobe once and it builds your entire catalog as you go.* Five minutes. You never have to do it again.
Fully buildable · YOLO on-device via TFLite
Recommended setup
Phone camera
Point your camera at any garment. AI identifies and logs it on the spot. Good for adding new pieces as you buy them.
Fully buildable · Flutter + TFLite
Add any time
Photo import
Already have photos of your clothes? Drop them in. ColorSense processes them server-side and populates your catalog without you lifting a finger.
Fully buildable · Server-side
Bulk import
Spectral scan
The sensor reads the actual spectral signature of the fabric — twelve channels, violet to near-infrared. Navy versus black. A camera guesses. This doesn’t.
Core hardware · Already in dev
Highest precision
*Walk-through scan requires the full ColorSense system. Not available on ColorSense.
**Camera-free cataloging also available via NFC/RFID for restricted environments. Learn about ColorSense for restricted environments →
How we build things
We looked at every approach. Here’s what we kept, what we cut, and why.
We explored
Ambient closet sensing
Leave a sensor in your closet and let it learn passively. Compelling in a pitch. In reality: clothes face edge-on, lighting is terrible. It fails the first time a real person tries it.
Cut → replaced with the 5-minute walk-through
We explored
Mandatory item scanning at setup
Ask users to scan every item before the system is useful. We heard “I’m out” from users, investors, and our own team. So we made scanning one of four options, not a requirement.
Kept as an option → never as a gate
We explored
Always-on cloud AI updates
Sounds great until it’s 6am, the wifi is flaky, and your device is staring at a spinner. The core function must work every single time. Cloud features are an enhancement, not the foundation.
Cloud as enhancement → never as dependency
KISS.
Keep it simple. Every feature got the same test: does this work reliably at 6am in a real person’s dark closet, the first time, without explanation? If the answer was “mostly” or “usually” — it didn’t ship.
ColorSense

Color & fabric.
One button.
Any language.

ColorSense does one thing: place a fabric on the sensor, press the button, hear the color. No app, no account, no wifi. It works in the dark because the sensor reads the fabric directly — twelve spectral channels that a phone camera doesn’t have.

It also arrives already knowing things — style rules baked in at the factory: navy and black fight, brown and black are different temperature darks, a neutral base with one accent works everywhere. At $349, this is the device that should have existed ten years ago.

ColorSense →
×
Navy + Black
Both dark, different undertones — they fight.
Navy + Gray
Shared cool tone, clean contrast. Always works.
×
Brown + Black
Warm dark meets cool dark — looks unresolved.
Navy + Tan
Classic contrast pair. The work is done for you.
×
Red + Pink
Two warm tones competing.
Brown + Tan
Tonal dressing — intentional, editorial.
×
White + Cream
Looks like a mistake, not a choice.
Black + White
Maximum contrast. Never wrong.
×
3+ distinct colors
Too many signals. The outfit loses gravity.
Neutral base + one accent
The formula that reads right in every culture, every room.