Jev

Jev

A structured decision model that turns text or JSON states into typed judgments—Choice, Score, or Noul—with probabilities.

freemium
Last Updated 9/29/2026

About Jev

Jev is a managed decision model from TypeSafe AI and the service offered by the JevModel website. You give it a state (text or JSON) and a focused question, and it returns a typed answer that software can directly check: one of three forms—picking one option from a set (Choice), a score on a scale (Score), or the probability that a statement is true (Noul). It does not write replies or make final decisions for your application: you can test its judgments on your own examples, choose action thresholds in your code, and send uncertain cases to human review. JevModel is an independent website and is not affiliated with TypeSafe AI.

Best for

Developers who need structured decisions on choices, ratings, or probability judgments inside their own software
Teams that want to set action thresholds in code and route uncertain results to human review
Individual users who want to try things out with the free quota before deciding whether to purchase capacity

Best use cases

Have the model pick one option from a set within your application, with a probability attached

When you need the model to score on a scale, send the state and question to Jev to get a Score result

When you need the probability that a statement is true, and decide subsequent actions in code by threshold

Instead of executing uncertain judgments directly, flag them and route them to human review

Features

Three judgment forms

Provide a state and a question, and get back one of three typed answers: a choice (Choice), a score on a scale (Score), or the probability that a statement is true (Noul).

Typed answers

The returned answer is a typed result your software can directly check, not a piece of reply text.

Playground and API

Call it from the web Playground or via the API (POST /api/v1/systemone); both share the same token balance.

Pros

  • Returns typed, structured answers that software can directly check and use
  • Three judgment forms (choice, score, probability) cover different types of decision problems
  • The Playground and API share the same balance, making it easy to try before integrating
  • Tokens never expire, and failed requests are not charged

Cons

  • Jev does not write replies or make final decisions on behalf of your application; callers still need to set their own thresholds and handle uncertain results
  • The free quota is counted by UTC day, with at most five successful free runs per day

Pricing Plans

Starter

$4.90一次性购买
  • 8M 代币,一次性购买
  • 每 1M 输入代币 $0.61
  • 按每次 500 输入代币估算约 16,000 次请求
  • 代币永不过期
  • Playground + API 共用一个余额
  • 失败的请求不收费
Most Popular

Builder

$19.90一次性购买
  • 50M 代币,一次性购买
  • 每 1M 输入代币 $0.40
  • 按每次 500 输入代币估算约 100,000 次请求
  • 代币永不过期
  • Playground + API 共用一个余额
  • 失败的请求不收费

* Pricing information is for reference only. Please refer to the latest announcement on Jev's official website.

FAQ

The free plan includes 5 successful free runs per day (UTC) plus 100K free tokens; the tokens never expire, the Playground and API share a single balance, and failed requests are not charged.
JevModel is an independent website and has no affiliation with TypeSafe AI.
Create a key in account settings, then call POST /api/v1/systemone.

Tool Information

DeveloperJev
Release Date9/29/2026
Pricingfreemium