OmniStudio
Local-first desktop AI workbench for llama.cpp, vLLM and SGLang that includes a JEV typed-decision page and /v1/systemone gateway.
Overview
OmniStudio is a local-first desktop workbench, documented in Chinese, for managing and serving local models with llama.cpp, vLLM and SGLang. It bundles chat, voice, image, video, music, OCR and translation apps, a prompt library, small utility apps, a local knowledge base with RAG, shared agent memory, skills management and an inference benchmark tool.
One page is dedicated to JEV typed decisions: noul, choice and score questions that return probability distributions and confidence rather than text, making classification, tagging and scoring comparable and thresholdable. You can run locally (it installs laya-mlx, downloads weights and starts the model on the same page) or connect to the cloud. Five bilingual examples cover ticket assignment, resume scoring, content guardrails, retrieval reranking and intent routing. Its gateway exposes /v1/systemone matching TypeSafe field by field, so the official SDK works by changing TYPESAFE_BASE_URL and TYPESAFE_API_KEY.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Choice | Which team should this ticket go to? | — |
| Score | How well does this resume match? | — |
| Yes / no | Does this content violate the guardrail? | yes · no |
How it works
- Choose local (laya-mlx) or cloud on the JEV page.
- Enter state and questions, or load a sample.
- Inspect per-question distributions.
- Call the local /v1/systemone gateway from the official SDK.
Strengths
- Local decisions with a TypeSafe-compatible gateway.
- Ready-made bilingual examples.
- Part of a broader local AI toolkit.
Caveats
- Large desktop app; JEV is one feature.
- Documentation is in Chinese.
- Local mode relies on laya-mlx (Apple Silicon).
Getting started
Download the desktop app from the project releases and open the JEV typed-decision page.At a glance
- Type
- Service
- Status
- beta
- Runtime
- desktop app, local gateway
- Has code
- yes
- License
- MIT
Use cases and integrations
How curated lists describe it
Local-first desktop workbench for llama.cpp, vLLM and SGLang that also serves `/v1/systemone` through its own gateway, with the contract written down field by field. Chinese.
MrJev/awesome-jev
Frequently asked questions
Can I use the TypeSafe SDK against OmniStudio?
Yes, change TYPESAFE_BASE_URL and TYPESAFE_API_KEY to point at its gateway.
Which local model does it use for JEV?
It installs laya-mlx and downloads weights from the JEV page.
What else does it do?
Chat, voice, image and video generation, OCR, translation, RAG and benchmarking.
Related projects
Go harness for System One models: hosted Jev, open Kev in-process via llama.cpp, and Cloudflare Clef, with typed helpers and a CLI.
ServicebetaGoTypeScript library for one-line local decisions with open models (kev-0.6b, kev-4b, open-jev), no API keys or cloud.
ProjectbetaTypeScriptSwift 6 bridge that runs System One decision models behind Apple's Foundation Models API, on device with Laya or via hosted endpoints.
ProjectbetaSwiftZero-dependency npm package that runs Laya natively or as WebAssembly in Node, Bun and browsers, drop-in for Jev's /v1/systemone.
ProjectbetaJavaScriptRuns the open LAYA decision model under Bun with native ONNX inference behind a TypeSafe-compatible API, plus an English/Persian playground.
LibrarybetaTypeScript- ollaya★ 1,305
Rust daemon and CLI that pulls open decision models like Laya and serves them locally behind a TypeSafe-compatible /v1/systemone.
ServicebetaRust