Jev AI Tutorials: Learn the Judgment Model by Doing

Bilingual tutorials, runnable case studies and version tracking for Jev 1.13 — the System One judgment model by TypeSafe AI that answers judgment, choice and scoring questions with typed output and confidence scores. Not a chat model.

Start: your first API call → What is Jev AI? →

One support message, three steps

01

Customer: my parcel has not arrived

02

Jev picks: logistics issue

answer: logistics · confidence: 0.93

03

System: route to logistics queue

No free-text reply. A typed decision your code can act on.

See what a Jev response looks like

Pick a question type, run the demo, and compare the typed structured output with the equivalent chat-model answer. All responses are preset example fixtures — no network request is made.

Example input

Run demo →

Make it real

Send the same question to Jev with your own API key:

curl https://openrouter.ai/api/v1/chat/completions \
  -H 'Authorization: Bearer $OPENROUTER_API_KEY' \
  -H 'Content-Type: application/json' \
  -d '{"model":"typesafe/jev-1.13","messages":[{"role":"user","content":"Is this comment spam?"}]}'

Example fixture — confirm the exact model slug and fields on the official docs / OpenRouter model page. · Full tutorial

Start with a guide

All guides →

Browse scenarios

All scenarios →

Case studies (full code)

All cases →

Version tracking

What changed in Jev 1.13 and how to migrate.

All versions →