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

typesafe-jev-playground

BYOK practice workbench that emulates Jev's Noul, Score and Choice workflows with ordinary LLMs, or calls native Jev if you have access.

Servicebeta
View on GitHub ↗

Overview

TypeSafe Jev Playground is an open-source web workbench for learning Jev concepts: state, atomic questions and the Noul, Score and Choice primitives. By default it runs in LLM practice mode, using a model you already have a key for (Groq, OpenAI, Anthropic, Gemini or local Ollama) to emulate Jev-style evaluations, so developers without Jev access can draft questions and rubrics.

The README is explicit that practice mode is not Jev and does not run TypeSafe's model; a native mode calls the real Jev API for users with credentials. It is built with TypeScript, React, Vite, Tailwind and Express, with Vitest tests. It suits people preparing Jev integrations before getting access, with the understanding that emulated probabilities will differ from Jev's.

What the decision model decides

TypeQuestionAnswers
Yes / noUser-defined yes/no questionyes · no
ScoreUser-defined rubric score—
ChoiceUser-defined choice—

How it works

  1. Choose LLM practice mode or native Jev mode.
  2. Enter state and questions.
  3. Run the evaluation and inspect answers.
  4. Refine questions and rubrics before using real Jev.

Strengths

  • Lets developers practice without Jev access.
  • Clear separation between emulation and real Jev.
  • Supports many LLM providers including local Ollama.

Caveats

  • Practice mode results are not Jev results.
  • Emulated probabilities are not calibrated like Jev's.

At a glance

Type
Service
Status
beta
Language
TypeScript
Runtime
React/Vite web app + Express server
Has code
yes
Has tests
yes

Use cases and integrations

How curated lists describe it

  • A BYOK playground for practicing TypeSafe's Jev — the first System One model — with your own LLM or native Jev API access. Experiment with Noul, Score, Choice, custom state, and structured decisions.

    logicrw/awesome-jev-projects

  • A BYOK playground for practicing TypeSafe's Jev — the first System One model — with your own LLM or native Jev API access. Experiment with Noul, Score, Choice, custom state, and structured decisions.

    hellogumbo/awesome-jev

Frequently asked questions

Is this the real Jev?

Only in native mode; the default practice mode emulates Jev with a general LLM.

Who is it for?

Developers learning or prototyping Jev workflows, especially without Jev access.

Which LLMs can it use?

Groq, OpenAI, Anthropic, Gemini or local Ollama.

Related projects

  • Building with Jev★ 155

    Agent skill for writing and improving programs that call Jev: question design, state structure, thresholds and diagnosing wrong answers.

    Agent skillproduction
  • typesafe.pro

    Server behind api.typesafe.pro, a free anonymous System One-compatible endpoint that forwards to TypeSafe on the operator's key.

    Serviceproduction
  • jev-agent-toolkit

    Portable Agent Skill teaching coding agents to use Jev as a bounded judgment layer, plus an optional MCP bridge.

    MCP serverbetaTypeScript
  • jev-anything

    Independent agent skill that makes a coding agent design, build, test and tune a bounded Jev decision layer with a contract first.

    Agent skillbeta
  • jev-architect

    Agent skill that guides you in finding, designing and testing Jev decision loops, including when to keep logic in code.

    Agent skillbeta
  • jev-builder

    Browser form for building Jev requests from pasted text without hand-writing JSON, then rerunning to see how steady the answers are.

    ProjectbetaJavaScript