Jev in production › Model and agent routing

Design OS Generative UI

Generative UI engine routes layout decisions to Jev instead of streaming LLM tokens; author claims responses of 6.5-25ms versus seconds-long token streaming (author).

Open on GitHub ↗

6.5-25msmeasured against a baseline, as published by the source
Use
Model and agent routing
Industry
Developer tools
Form
Open-source tool
Stage
Beta
Listed
2026-09-26
Found via
github
Repository
jangtrinh/design-os-generative-ui
Stars
3
Forks
0
Last push
2026-09-26
Language
TypeScript
License
none stated

The README opens with

Sub-50ms Real-Time Generative UI Engine Powered by Laya-MLX (Apple Silicon Local Edge) & TypeSafe JEV (Cloud API) with the Catalog Pattern (Zod).

⚡ Live demo recorded at 60fps in the interactive playground: instant sub-50ms layout snapping without LLM token streaming delay.

Traditional Generative UI tools (v0, Claude Canvas, ChatGPT Canvas) ask a large language model to stream raw JSX or HTML tokens. In production applications, this introduces severe bottlenecks:

Badge

For the project's own README, linking back here:

Listed in Jev in production

Also used for model and agent routing