JEV / DECISION ENGINEERING

Turn a fuzzy judgment into a reliable software step.

Start with a real task. Design the state, typed questions, and human fallback. Examples follow official documentation; performance numbers always include their source and scope.

Start learning Design a decision

One support ticket runs through the course. Learn the input and questions, inspect the official API, then design a fallback using labeled examples. About 60 minutes of reading; practice time varies.

STATE→QUESTION→DECISION→CODE

INPUT / 01OUTPUT / 02

My parcel says delivered, but nothing arrived.

↘ CHOICE · support_team

shipping0.82

billing0.11

other0.07

This is a design draft, not an actual Jev answer or accuracy test.

Start here

Choose your next step

Learn the boundary, make a first request, then check outcomes on examples.

0101 / COURSE

Jev beginner course: from first principles to a testable workflow

Learn Jev's boundaries, state, the three question types, API answers, confidence, and validation in order. Practice before getting a key.

0202 / BUILD

Your first API call

Use the official state, model, and questions structure to make a first call and inspect its answer.

0303 / APPLY

Decision Design Lab

Describe a narrow task and get an editable, copyable request draft. This tool does not call Jev or validate model performance.

USE CASES

Find a use case by task

Each example separates the input, Jev's bounded decision, and the action owned by code or a person.

Use cases

Choice

Jev learning resources: three independent sites mapped

Use Playground for real runs, Manual for downloadable programs, and Learn Jev for patterns and reliability. Each entry links to the original and to a matching primer here.

Topic directory

SOURCE FIRST

Evidence and limits

We have not run a Jev API benchmark. Independent learning site. Not affiliated with TypeSafe AI.

Official docs Public cases