Jev AI API

Jev AI API for structured decisions

Jev AI API turns application state into typed decisions, probabilities, and confidence signals your backend can use for routing, scoring, guardrails, and review.

Try it in Playground Read the API docs

Decision requestLIVE

INPUT STATE

Three deploys have failed and production is returning 500s.

choice

Classify the next action

route

confidence

Probability-backed signal

0.94

needs_human

Escalate when risk is high

true

One state. Multiple typed questions. A response your code can inspect.

API surface: REST
Question types: 3 types
One request: Parallel
Final action: Your code

Build with the API

A predictable request shape for production

Start with a real state, define the smallest useful question, and pass the structured result into the service that already owns the next action.

POST /v1/systemone

A minimal server-side request to the Jev AI API

curl -X POST https://thejevai.com/v1/systemone \
  -H "Authorization: Bearer $JEV_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "typesafe/jev-1.13",
    "state": "Three deploys failed and production is returning 500s.",
    "questions": {
      "needs_human": {
        "type": "noul",
        "instructions": "Does this need human review now?"
      }
    }
  }'

// Read data.result.answers.needs_human.noul
// and data.creditsUsed in your service

One request, clear output

A request shape your backend can own

Start with state, choose a Jev model, and describe the questions your product needs answered. The result comes back in a predictable shape for routing, scoring, guardrails, and review.

Typed questions

Use Choice, Score, or Noul to define the answer space before the request reaches the model.

Managed API keys

Create and rotate keys from your Jev workspace, then keep them in server-side environment variables.

Usage you can see

Inspect credits used and request usage so you can estimate the cost of a workflow before expanding it.

Backend-ready results

Read answers by question ID and let your own code decide whether to route, queue, block, or ask a person.

Parallel questions

Ask several related questions against the same state without chaining separate calls for every signal.

Explicit ownership

Use Jev for bounded judgment while your service owns authentication, thresholds, retries, and final actions.

From input to action

The API fits the decisions already hiding in your code

Jev handles the judgment in the middle. Your existing services still choose the model, apply policy, and decide when a person needs to step in.

Validate the question

Use Jev Playground with a real low-risk state and refine the question until the output is useful.

Create an API key

Create a key in your Jev workspace and store it outside source code, browser bundles, and logs.

Keep the action in code

Use probabilities and confidence as signals, while your permissions, thresholds, and business rules own the final action.

Jev AI API questions

Understand the API before you connect it

What is the Jev AI API?

The Jev AI API turns a state plus typed questions into structured decisions, probabilities, and confidence signals that your backend can route, score, review, or use in guardrails.

Which question types does the API support?

Jev supports Choice for classification and routing, Score for ordered ratings, and Noul for yes-or-no decisions. Multiple questions can be evaluated against the same state in one request.

Does Jev execute the final action?

No. Jev provides a bounded judgment. Your application remains responsible for authentication, thresholds, retries, permissions, audit logs, and the final action.

Jev AI platform

Move from a tested question to a real workflow

Start with one controlled decision.

Jev performs the bounded judgment. Your service remains responsible for authentication, input minimization, retries, rate-limit handling, thresholds, audit logs, and the final action.

Create an API key View plans