Local AI, running on a local workstation.

No cloud AI. No big-tech APIs. Pick a demo — every response and every render runs on a local workstation, served straight to this booth. This is the ODS stack, by Osmantic.

Chat with the model

Open conversation with Qwen3-Coder-Next running on a local workstation. No cloud, no API.

Start chatting

Generate a video

Type a scene, get a short video clip from LTX-2.3 running on a local workstation. Take it home with a QR code.

Make a video

Jailbreak a local LLM

A local LLM is playing a jaded VC with one rule: never admit your startup is any good. Can you jailbreak it into breaking character and calling your idea worthy? No one has yet.

Break it

Race the cloud

Same prompt, two streams, one race. A local workstation against a frontier cloud datacenter — watch them answer in real time.

Start the race