Skip to content
Metatext Decide — typed decisions for your agents, with a probability on every answerGet early access
Metatext
Sign Up

laya-system-one

Zero-dependency npm package that runs Laya natively or as WebAssembly in Node, Bun and browsers, drop-in for Jev's /v1/systemone.

ProjectbetaApache-2.0
View on GitHub ↗

Overview

Laya System-One packages the Laya decision model so it runs entirely on your machine in Node.js, Bun or the browser with no Python, PyTorch, API keys or cloud calls at inference time. A single npm install selects the right native engine for your platform, or WebAssembly in the browser, and the roughly 324 MB model is fetched once and cached. The package has no dependencies.

It can run as a local HTTP service via npx with optional bearer authentication, exposing a drop-in POST /v1/systemone so existing Jev clients work, or in-process for zero network overhead. The README reports about 40 ms per question on a warm engine across shipped platforms and support for over 100 languages. It is a community project built on Convai Innovations' Laya, credited in the README, and suits JavaScript developers wanting offline typed decisions on edge servers, intranets or in the browser.

What the decision model decides

TypeQuestionAnswers
ChoiceWhich category or team fits this message?(inferred)—
Yes / noAny yes/no question about the state—
ScoreAny ordered score question about the state—

How it works

  1. npm install laya-system-one
  2. Run npx laya-system-one --port 8080 or Laya.load() in-process
  3. Model downloads once and is cached
  4. Send state and typed questions; receive probabilities

Strengths

  • Zero dependencies and fully offline
  • Runs in Node, Bun and browsers
  • Drop-in Jev endpoint
  • Optional bearer auth

Caveats

  • Community packaging of Laya, not official
  • Accuracy is Laya's, below hosted Jev on many tasks
  • 324 MB model download

Getting started

npm install laya-system-one, then npx laya-system-one --port 8080

At a glance

Type
Project
Status
beta
Language
JavaScript
Runtime
Node, Bun (native) and browser (WASM/WebGPU)
Has code
yes
License
Apache-2.0

Use cases and integrations

How curated lists describe it

  • Self-contained, ultra-fast System 1 decision engine with WebGPU/WASM acceleration, 100% wire-compatible with TypeSafe Jev (/v1/systemone).

    hellogumbo/awesome-jev

Frequently asked questions

Does it need Python?

No, it ships native and WASM engines with no dependencies.

Can I use my Jev client?

Yes, the HTTP service exposes POST /v1/systemone.

How fast is it?

About 40 ms per question on a warm engine, per the README.

Related projects

  • gojev

    Go harness for System One models: hosted Jev, open Kev in-process via llama.cpp, and Cloudflare Clef, with typed helpers and a CLI.

    ServicebetaGo
  • jev

    TypeScript library for one-line local decisions with open models (kev-0.6b, kev-4b, open-jev), no API keys or cloud.

    ProjectbetaTypeScript
  • jev-foundation-models

    Swift 6 bridge that runs System One decision models behind Apple's Foundation Models API, on device with Laya or via hosted endpoints.

    ProjectbetaSwift
  • OmniStudio

    Local-first desktop AI workbench for llama.cpp, vLLM and SGLang that includes a JEV typed-decision page and /v1/systemone gateway.

    Servicebeta
  • usejev

    Runs 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