Jev AI

TypeSafe System One model

Jev AI

LIVE

Available now in the playground and API.

Ask questions about any text. Get probabilities, not prose.

Try Jev AI free Jev AI API
  • 20 welcome credits
  • ~0.5 s answers
  • Yes/No · Choice · Score

Jev AI playground

About GPT-6 Luna · Use it through the API · OpenAI’s Decisions API. Text, JSON and up to 4 images per run; 3× model rate.

3 credits per run, or actual input tokens × 3 when credits are insufficient. Output tokens are free. All model rates.

1 Input
Images optional0 / 4 · 1 KB of 8 MiB

PNG, JPEG or WebP · up to 4 MiB and 16 megapixels each. Images are placed before the text and are not saved in task history.

2 Questions
3 Answers
Run GPT-6 Luna to see answers

Build with Jev AI

Use Jev AI to judge answers, route requests and check retrieved context, each with your own rules.

Evaluation & quality

LLM as a Judge

Evaluate an answer for source support, relevance and quality using your own rubric.

  • Answer quality
  • Groundedness
  • Rubrics
Open tool & guide
Routing & decisions

LLM Router

Choose a model tier, handler or tool for each request, with confidence to fall back safely.

  • Model routing
  • Semantic routing
  • Tool selection
Open tool & guide
Retrieval & knowledge

RAG Evaluation

Check each retrieved passage for relevance, answer coverage and injected instructions before generation.

  • Context relevance
  • Reranking
  • Prompt injection
Open tool & guide

More Jev AI use cases: AI Agent EvaluationEntity MatchingClaude Code Rules CheckerMCP Tool RouterSupport Ticket TriagePrompt Injection DetectorLead QualificationSales Call ScoringJev JobsLive Web Context · browse all by category →

What is Jev AI?

Jev is TypeSafe's first System One model. Jev AI lets you try it and call it by API.

Typed answers

Numbers and labels your code can branch on.

Calibrated

Choice and Score include confidence; Yes / No returns a probability.

Many questions, one call

All answered in parallel.

Same API shape

TypeSafe-compatible requests.

New to the model? Read what Jev does and where its limits are, then follow a recorded support-ticket example before running your own input.

Three Jev AI question types

Yes / No

Probability that the answer is yes.

Choice

Picks one option and shows every probability.

Score

Rates the text on your own levels.

How Jev AI works

  1. 01Pick a case
  2. 02Edit questions
  3. 03Run Jev AI
  4. 04Ship with the API

Ready to integrate? Follow the API request and response reference, process a batch of text inputs, or set up Jevgrep for code search. Keep permissions and final actions in your application; use probabilities to decide when a person should review a result.

Jev AI pricing

Start20 welcome credits
All paid usageInput tokens only
Web · APICredits first · Tokens first
See plans

Jev AI FAQ

Read: what is Jev?

What is Jev AI?

Jev AI (jev-ai.pro) is an independent playground and API for Jev, the System One model from TypeSafe. You give Jev a piece of text and typed questions; it returns yes/no probabilities, choices and scores with calibrated confidence.

What is the Jev model?

Jev is TypeSafe's flagship model and the first System One model. It is trained for fast structured decisions instead of text generation: it evaluates many narrow questions about one state in parallel and reports how sure it is.

What is the Jev AI platform?

Jev AI brings TypeSafe’s Jev models into one workspace for playground, batch processing, and API access. The platform is independently operated, with subscriptions, token balances, and customer support managed by Jev AI.

How does the Jev API on this site work?

Create an API key on the Jev API page and send the same JSON you would send to TypeSafe (state, model, questions) to https://jev-ai.pro/api/v1/systemone. API calls use the Token account first; model multipliers apply, and insufficient tokens switch the entire request to model-priced credits.

How much does Jev AI cost?

Eligible new accounts get 20 welcome credits. Welcome and daily check-in gifts enter the Credit account; monthly plans and one-time purchases add input tokens to a separate Token account. Web tools use model-priced credits, then input tokens × model rate if credits run out. API calls use model-priced tokens first, then the model’s credit charge if tokens are insufficient. A request never charges both accounts. Output tokens are free.

Can Jev read images or PDFs?

Native Jev accepts text only: a string, a JSON object or an array. For images, audio and short videos, select Jev-Omni in the Playground. PDF upload is not supported; convert PDF content to text first.

Which languages does Jev support?

English is Jev's primary language and where it is most accurate. Other languages work but less reliably, so test on your own data and watch the confidence value.