Open· Last updated

Jev learning path

A “learning path” on jev.pro is a sequence of intents, not a bootcamp, certification, or TypeSafe onboarding product. You already heard of Jev / System One. This page gets you from curious to Pro in the NARRATIVE.md sense: production-minded craft — not a paid tier.

Independent unofficial field guide. Canonical APIs and model cards: docs.typesafe.ai. This site does not sell or issue keys.

Independent angle (cover, do not clone)

Win delta: pillar-ordered path (Explain → Integrate → Compose → Compare → Operate) plus dual-stack checkpoints. We do not mirror a rival handbook IA or a “Jev course” slug tree.

Question and context

Start with the entity: Jev is TypeSafe’s flagship System One decision model. You send a small state and typed questions (Choice / Score / Noul); your code acts on probabilities and (for Choice/Score) confidence. Official guidance: a good question is a snap decision a knowledgeable person could make in a few seconds.

If you need the mental-model essay first, use What is Jev (open cut) and the explainer what-is-jev. This URL is the order to read and practice, not a second what-is.

Core concepts (week-shaped, not calendar promises)

Stage Pillar Exit gate Start here
1 Explain You can say what Jev is not (chatbot, generator, official docs) What is Jev, how Jev decides, primitives
2 Integrate One door, one pinned id, one traced call first API call, channels, version pinning
3 Compose Code owns side effects; LLM drafts after a decision intent routing, guardrails
4 Compare You know when not to use Jev vs prompt-only, vs classifiers
5 Operate Floors, gold JSONL, fail-closed playbook, evaluation, best practices

Do not skip jaggedness: TypeSafe documents jev-1.13 as weak on arithmetic, dates, noisy state, and adversarial text. Keep counts in code. Vendor speed multiples and the $0.042 / M input list price are vendor claims — confirm on the models page.

Practical approach

  1. Recreate the dual-stack snippet on what-is-jev against your key (console or a documented gateway — not us).
  2. Pin jev-1.13.0 (versioned id) once you touch thresholds. Log response.model.
  3. Freeze one questions file. Add a 20-row gold set before a second workflow.
  4. Read common mistakes and anti-patterns before you invent a 40-question “brain.”
  5. Only then open a vertical pack (use cases).

Risks and trade-offs

A path that starts at “recipes” without primitives produces prompt-shaped Choice criteria and silent alias drift. A path that only reads official docs without operate loops ships Playground screenshots. We will not invent study-time or pass-rate metrics.

What this page does not claim

FAQ

Do I need every cluster page? No. One intent per URL. Follow the table until the exit gate is true, then stop.

Python or TypeScript first? Whichever you ship. Keep the JSON body identical if both exist — see the implementation checklist.

Where do keys come from? TypeSafe’s console or a documented marketplace door. jev.pro does not issue keys.

Disclaimer

This is an independent unofficial site and is not affiliated with TypeSafe AI; official documentation is available at https://docs.typesafe.ai. Never treat jev.pro as TypeSafe official documentation. We do not sell, issue, or proxy API keys.

Open-cluster pages are independent field-guide notes. Replicas and third-party interfaces mentioned anywhere on jev.pro are not Jev and not endorsed. Hub: Open. Siblings: quick reference, playbook, what is Jev. Canonical: https://docs.typesafe.ai.

Sources

Public TypeSafe or adjacent documentation only. No private claims.