On-device card vision
Track, extract and classify every card in frame, in real time, on the device holding the camera.
soft body. hard tech.
JellyTek
Edge-native computer vision and market data for trading card games. Our models run on the device in front of you — no cloud round-trip.
soft body.hard tech.
Computer vision · Market data · Edge deployment
Scroll
01 — Who we are
A computer-vision company
A computer-vision company
that works on cards.
Small team, deep specialism.
We came out of automotive AI — cameras that have to work on every frame, in bad light, on a tiny power budget, without phoning home. We pointed that discipline at trading cards.
Our vision
Every card, on every device, recognised instantly and priced honestly — without sending a picture to somebody else's cloud.
02 — Expertise
What we
01
02
03
04
What we
specialise in.
Four things, all pointed at the same problem.
Embedded & NPU deployment
Built for mobile silicon and low-cost SoCs, not just a workstation GPU.
Synthetic data pipelines
We generate the training set rather than annotate it.
Market data & pricing
Pricing and history wired to the same recogniser that reads the card.
03 — Products
Two products,
Two products,
one foundation.
A recognition engine, and the market data that sits on top of it.
04 — Team
The people
The people
behind it.
Deep embedded machine-learning experience across the team.
johnny
CTO · Engine
Builds the engine — the runtime, the models, and everything that has to survive contact with real hardware.
tess
App Design
Designs the apps that sit on the engine, and the way people actually use them.
james
Business
Partnerships, customers, and getting the thing in front of the people who need it.
05 — Demo
Let us show you
Let us show you
the real thing.
marketing@jellytek.net
0439 505 062
The short version is on this page. We walk through the rest — live, on real cards — in a demo.