Bud-Decision-Studio
LM Studio-style desktop app and local server for open Jev-like decision models, with a playground, templates, API and LoRA training.
Overview
Bud Decision Studio, from Bud Ecosystem, runs open decision models on your own computer. You give a model a situation and questions and get a calibrated probability for every answer in milliseconds; it describes itself as the LM Studio of Jev-like System One models. It installs with one line or as native packages for macOS, Windows and Linux.
It includes a models table of eleven open models (such as Laya, Kev and Jev-Omni) with what each is good at, size, input types and published results next to Jev, one-click download and load, a Playground that charts every answer, versioned templates with decision history, a Train page for LoRA fine-tuning on your own labelled decisions, and a local /v1/systemone-compatible API with examples in curl, Python, JavaScript and the official TypeSafe SDKs. Documentation lives in the repository's site folder. List summaries note no licence file yet.
What the decision model decides
| Type | Question | Answers |
|---|---|---|
| Yes / no | Does it need an answer today? | yes · no |
| Choice | User-defined choice questions in the playground | — |
How it works
- Install and pick a model
- Ask typed questions in the playground or via API
- Save templates and review decision history
- Optionally fine-tune with LoRA on labelled decisions
Strengths
- Local and private
- Works with TypeSafe SDKs
- Built-in fine-tuning
Caveats
- No licence file yet
- Some models need an NVIDIA GPU
Getting started
curl -fsSL https://raw.githubusercontent.com/BudEcosystem/Bud-Decision-Engine/main/get.sh | shAt a glance
- Type
- Service
- Status
- beta
- Runtime
- desktop app + local server
- Has code
- yes
Use cases and integrations
How curated lists describe it
a cross-platform desktop app (macOS, Windows, Linux) that runs `laya`, `kev`, `laya-typed-decisions` etc locally behind the Jev API (`POST /v1/systemone`), with a Playground, versioned templates with decision history, and a Train page for LoRA fine-tuning on your own labelled decisions.
yibie/awesome-jevOpen-source, cross-platform desktop runtime and playground from Bud Ecosystem for running, evaluating, managing, and training Jev-like decision models locally, with a `/v1/systemone` API. Supports 10+ decision models.
AnotiaWang/awesome-decision-modelsBud Decision studio is an easy to use cross platform desktop application &…
wh000wh000/awesome-jev-liveBud Decision studio is an easy to use cross platform desktop application & serving system for Jev like models locally.
hellogumbo/awesome-jevDesktop app and local server for open Jev-style decision models on macOS, Windows and Linux: pick a model, ask typed questions, read every answer as a chart, or call it over a `/v1/systemone`-compatible API. No licence file yet.
MrJev/awesome-jev
Frequently asked questions
Which models does it run?
Eleven open models such as Laya, Kev and Jev-Omni.
Can my TypeSafe SDK code use it?
Yes, it exposes a /v1/systemone-compatible API.
Can I fine-tune?
Yes, LoRA training on your own labelled decisions.
Related projects
Docker service exposing local Laya, Von, a CLM head or hosted Jev through one TypeSafe-compatible HTTP and MCP endpoint.
MCP serverbetaPythonConverts open Jev reproductions (Laya, Kev, NanoJev, PlayJev) to INT8 ONNX for offline CPU decisions at about 15.6 ms per question.
ServicebetaPythonNative Go runtime serving Gemma 4 12B as a Jev-like decision model with hand-tuned PTX kernels and no CGo or CUDA toolkit.
ProjectbetaGoGo harness for System One models: hosted Jev, open Kev in-process via llama.cpp, and Cloudflare Clef, with typed helpers and a CLI.
ServicebetaGoGo gateway exposing the Jev /v1/systemone contract over ordinary LLM backends by reading single-token label logprobs.
ServicebetaGoRuns Intern-Decision-0.8B on Apple Silicon behind a local /v1/systemone endpoint that accepts screenshots for agent decisions.
ServicebetaPython