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

typesafe-jev-decision-studio

Interactive web workbench for exploring and benchmarking Jev's Choice, Noul and Score via OpenRouter, with threshold transfer curves.

Projectprototype
View on GitHub ↗

Overview

TypeSafe Jev Decision Studio is a Node-based interactive workbench for exploring Jev's three primitives through OpenRouter (typesafe/jev-1.13). It shows Choice classification over large option sets (the README's example uses 95 categories of self-hosted software), Noul binary checks with probability transfer curves and user-set thresholds, and Score ratings with full distributions across rubric tiers.

It is aimed at developers evaluating whether Jev fits their classification or verification tasks. The README cites 400 to 600 ms end-to-end latency via OpenRouter. Some of its language ("zero hallucinations", "mathematically calibrated") is promotional and should be read as the author's framing rather than measured claims. It requires Node 18.18+ or 20+ and an OpenRouter account with credit, and the model can be changed through an environment variable.

What the decision model decides

TypeQuestionAnswers
ChoiceWhich of 95 project categories fits this description?—
Yes / noBinary verification with a configurable thresholdyes · no
ScoreWhich rubric tier does this input fall into?—

How it works

  1. Configure an OpenRouter key.
  2. Enter state and a Choice, Noul or Score question.
  3. View the distribution, transfer curve or tier breakdown.
  4. Adjust thresholds interactively.

Strengths

  • Visual exploration of large option spaces and thresholds.
  • Uses OpenRouter, so no TypeSafe early-access key needed.

Caveats

  • Some claims are promotional.
  • Requires OpenRouter credit.

Getting started

git clone https://github.com/Romain-Jochum/typesafe-jev-decision-studio.git && npm install, set OPENROUTER_API_KEY in .env

At a glance

Type
Project
Status
prototype
Language
TypeScript
Runtime
Node web app
Has code
yes

Use cases and integrations

How curated lists describe it

  • Fast, calibrated System One decision platform powered by TypeSafe Jev via OpenRouter. Sub-second logprob scoring, transfer curves, zero hallucinations.

    logicrw/awesome-jev-projects

Frequently asked questions

How does it reach Jev?

Through OpenRouter using the typesafe/jev-1.13 model.

What can I explore?

Choice over many options, Noul with transfer curves and thresholds, and Score rubrics.

What latency does it report?

400 to 600 ms end to end via OpenRouter, per the README.

Related projects

  • jev-vs-laya

    Korean study notes and TypeScript mock-ups of Jev concepts, with a local Laya versus GPT-5 comparison and a calibration toolkit.

    ProjectresearchTypeScript
  • Jev calculator

    Calculator with no arithmetic in its code: Jev picks each answer character from 13 options, exposing its distributions and jitter.

    ServiceprototypeTypeScript
  • myJEV

    Learning project that builds choose-an-answer decision models from scratch and from Qwen, with downloadable 0.8B and 4B releases.

    LibraryresearchPython
  • jev-survey

    Preprint survey of Jev-style typed decision models: their research roots, adaptation costs and what current evidence supports.

    Modelresearch
  • Confidence

    Official TypeSafe docs page explaining how Jev's confidence is derived from the probability distribution and how to set thresholds.

    Articleproduction
  • jev-pipeline-skill

    Agent skill and CLI that chains Jev question design: find decision calls, check fit, draft, lint, calibrate thresholds and plan a cascade.

    Agent skillprototypePython