FURCATE
BASE

Deploy the GPU/CPU Furcate variant to accelerated edge hardware. Perception-heavy workloads — vision, audio, multi-stream — on-site, with a desktop UI when you want one. Where rich modality lives in the world model.

Furcate Base · reference

Specification

Hardware	Jetson Thor · Jetson Orin (Nano · NX · AGX) · GPU mini-PC · Hailo-10H accelerator class
RAM	8 – 64 GB · with NPU or discrete accelerator
Deploy	CLI · desktop
OS	Linux · Windows
Modality	Rich-modality — vision, audio, multi-stream perception, on-device language.
Role	Perception-heavy workloads. Where modality lives. The Furcate Edge AI class.
Network	Wired Ethernet · Wi-Fi. Sits inside the mesh, often serves model-to-model requests from Micro and Small neighbours.
Identity	Issued a Furcate machine identity — a local DID — at first deploy.

Edge AI class · perception-heavy deployments

Start a deployment ────►All four nodes ────►