OEB
Olyxee Edge Box
Olyxee Edge Box

AI that runs
in the real world

Olyxee Edge Box (OEB) is our execution division — the hands of Olyxee. We take AI from the lab and deploy it into the real world.

AI-powered computer vision across real-world environments — object detection, pose estimation, anomaly detection, and scene analysis running on edge devices

Computer vision at the edge

Object detection, pose estimation, anomaly detection, and scene analysis — deployed across industrial, automotive, and public safety environments.

How OEB fits into Olyxee

Three layers working together to deliver trustworthy AI.

Olyxee

The Brain

Defines vision, conducts research, and builds core AI capabilities. Strategy and intelligence originate here.

OEB

The Hands

Deploys AI into real-world systems. Takes models from lab to production across edge hardware at scale.

Grysics

Quality Control

Verifies AI works correctly before and after deployment. The trust layer that ensures reliability.

Edge capabilities

Everything you need to run AI at the edge, built in.

Hardware Abstraction

Deploy to any edge device — NVIDIA Jetson, Raspberry Pi, Arduino, ESP32 — with a single unified pipeline. No device-specific rewrites.

Runtime Optimization

Automatic quantization, pruning, and compilation tailored to each target device. Maximum performance from minimum resources.

Fleet Orchestration

Manage thousands of edge devices from a single control plane. Over-the-air updates, version control, and rollback across your entire fleet.

Secure Execution

Encrypted model delivery, secure enclaves, and tamper detection. Your models stay protected from device to cloud.

Multi-Model Pipelines

Chain multiple models on a single device — preprocessing, inference, and postprocessing — orchestrated as one coherent pipeline.

Edge-Cloud Sync

Seamless data flow between edge and cloud. Local inference with cloud fallback, telemetry aggregation, and continuous learning loops.

Supported hardware

One pipeline, any device. OEB abstracts hardware differences so you write your model once and deploy everywhere.

NVIDIA Jetson

GPU

Raspberry Pi

SBC

Arduino

MCU

ESP32

MCU

Intel NUC

x86

Google Coral

TPU

Qualcomm RB5

SoC

BeagleBone

SBC

Deploy AI
at the edge

Get your models running on real hardware in minutes. OEB handles the infrastructure — you focus on the intelligence.

Engineer managing edge infrastructure in a server room