Organization repository on GitHub
OmniStudio 是一个本地大模型一体化桌面工作台，集模型市集下载、llama.cpp/vLLM/SGLang 三引擎推理管理，以及对话、语音合成、ASR语音识别、图片生成、视频生成、OCR 等多种大模型应用于一体，全程本地优先。
- Forks
- 43
- Last pushed
- Updated Oct 2026
Repo stats from the GitHub API, cached Oct 2026.Homepage
CommunityJev Tools & CLIs
OmniStudio
A Chinese-first desktop workbench for local models whose JEV page runs typed noul, choice and score questions against a local engine or a cloud key, alongside a gateway, RAG knowledge base and shared memory.
- Stars
- 197
- Language
- TypeScript
- License
- MIT
- Published by
- Community
- Author
- kunpengtalk
- Last verified
- Oct 2026
Used inJev for Software & DevTools
- Languages
- TypeScript
Highlights
- The JEV page switches between local and cloud engines and configures them in place, then shows the probability distribution behind each question.
- Protocol matches TypeSafe field for field, so the official SDKs need only the base URL and API key lines changed to hit the local gateway.
- The local path hosts its own engine on Apple Silicon for fully offline use at no cost; two other entry points are an agent tool and the gateway endpoint.
- Five bilingual examples load in one click, and you can write a sentence and have the chat model translate it into a request body.
Watch out
The README describes the JEV page in demo terms and does not document which weights the local engine loads, its context limits, or any accuracy comparison against the hosted service, and the linked systemone doc was not part of what was read. Cloud access sends requests to TypeSafe under your own key, and the knowledge base, memory and skills features handle real data with their own storage and exposure paths. Internal counts also disagree, with seven mini apps in the interface section and eight in the feature list. Copy reviewed against the repository README on 2026-10-04, not from running it.
Questions about this entry
Short answers drawn from the linked project or page.
The page is presented as free, because the local path runs a model you host and is described as fully offline with no charge. Switching to cloud access means using your own TypeSafe key, so your account billing applies there. The README does not state which weights the local engine loads or where they come from, so check the linked systemone doc before assuming the local path matches the hosted model.
That is the stated goal for the core request shape: the gateway uses the same field layout, and the official Python and JavaScript SDKs need two lines changed to the local base URL and key. The same gateway also exposes the systemone endpoint and the knowledge base over MCP, so agents can reach it without an SDK. Anything beyond the core request shape is documented only in a separate file.
A platform. Beyond the JEV page it ships a model marketplace over a model hub, a three-engine runtime covering llama.cpp, vLLM and SGLang with hot switching, chat, voice, image, video, music, OCR and translation apps, a hybrid retrieval knowledge base with citations, cross-agent long-term memory, and a skills library synced into coding tools. Using Jev here means accepting the workbench's layout for prompts, knowledge and memory.
The outbound paths are named per feature: cloud access uses your own TypeSafe key, the knowledge base can call an OpenAI-compatible embeddings endpoint and optional rerank endpoints from three named providers, and chat retrieval can hit Bing, DuckDuckGo or Tavily with sources injected into context. The local retrieval, memory store and local decision path are described as on-machine, but the README says nothing about retention or deletion for the cloud-dependent parts.
It is MIT-licensed and built by KunpengTalk (鲲鹏Talk); it is a local-first desktop workbench rather than an SDK or a hosted service.
llama.cpp by default, plus vLLM and SGLang behind one abstraction with hot switching, and it can also connect to any OpenAI-compatible endpoint.
It splits local run versus cloud access, configures the engine, shows each question's probability distribution, and ships five bilingual runnable examples such as ticket dispatch, resume scoring and intent routing.
More Jev resources like this
56GitHub stars
R.A.I.N. Lab is a local-first research environment. Four constrained perspectives investigate a question, quoted spans are checked against a local paper corpus, and experiments are pre-registered with host code, not a model, deciding pass, fail or inconclusive.
Jev Tools & CLIs#community#jevbest
Community
topherchris420/james_libraryRust
topherchris420/james_libraryRust38GitHub stars
Python memory filter that judges every stored memory against the current request with one calibrated yes/no per memory in a single Jev request, keeping everything above a threshold.
Jev Tools & CLIs#community#jevbest#catalog-import
Community
samdotmak/jev-recallTypeScript
samdotmak/jev-recallTypeScript30GitHub stars
Browser extension that filters your X timeline with a model you choose and pay for, using plain-language rules plus per-author and blocked-word controls.
Jev Tools & CLIs#community#jevbest#catalog-import
Community
tshmieldev/sharpTypeScript
tshmieldev/sharpTypeScript27GitHub stars
Pi extension that replaces many small tool calls with one typed TypeScript execute_tools call, type-checked before running in a sandbox.
Jev Tools & CLIs#community#jevbest#catalog-import
Community
boozedog/pi-codemodeTypeScript
boozedog/pi-codemodeTypeScript17GitHub stars
Pi extension that trims wasted context before each model call by deduplicating repeated tool results, stripping failed payload args, and letting the model summarize closed spans.
Jev Tools & CLIs#community#jevbest#catalog-import
Community
Davidcreador/pi-dcpTypeScript
Davidcreador/pi-dcpTypeScript4GitHub stars
Shell CLI that sends state plus noul, choice, and score questions to Jev in one request and prints probabilities, choices, and rubric positions for pipes and scripts.
Jev Tools & CLIs#community#jevbest#catalog-import
Community
y0usaf/typesafe-cliTypeScript
y0usaf/typesafe-cliTypeScriptFrom the Jev guides
Original write-ups on what this entry is for.
The request shape barely changes across surfaces; the surrounding contract does. When to call the API directly and when a gateway, a Worker binding, or a router entry earns its extra hop.
Jev Tools & CLIs#gateway#vercel#cloudflare
Read article
Taking a decision checkpoint out of Python: what the GGUF and browser WebGPU exports preserve, what they change, and the parity evidence the model cards publish.
Jev Tools & CLIs#open-models#local#onnx
Read article
Jev labels a record that already exists. Code moves, flags, or holds it. How the document, mail, and transfer tools keep the action out of the model.
Jev Tools & CLIs#classification#email#security
Read article
From the Jev community
Posts from builders shipping with Jev right now.