Model profile / Jev

Jev decision model: inputs, outputs and task fit.

This Jevai guide explains TypeSafe's Jev decision model (including Jev 1.14) and the Choice workflow available on jevai.pro. Give it text context and allowed options, then inspect the selected label, probabilities and confidence. Jevai is an independent site, not the model provider.

Illustration of ranked choices with a highlighted leading result and a separate human review route.
A leading choice, a distribution, and a separate review policy. Concept illustration, not an API result.
Model specifications

Jev model specifications and Playground limits.

The facts below describe the documented API. Jevai.pro is an independent guide and Playground; TypeSafe operates the model API.

Jev model specifications and Playground limits.
SpecificationJevIn your application
Model identifierjev-latestA moving alias. TypeSafe also documents jev-preview and pinned version IDs such as jev-1.14.0 and jev-1.13.0. Record response.model when evaluating a run.
Verified alias targetjev-1.14.0; documentation checked 2026-10-28Both jev-latest and jev-preview resolve to jev-1.14.0 (released 2026-10-14); jev-1.13.0 remains available as a pinned version.
Input modalityText (plus opt-in API images)This Playground accepts text and CSV, TSV, or TXT tables. Direct TypeSafe API calls on Jev 1.14 and 1.13 also support opt-in image assets.
Request structuremodel + state + questionsThe Playground maps Context, Question and Choices into a Choice question in the official request.
Choice responsechoice + probabilities + confidenceUse the selected label and inspect the distribution. The TypeSafe API also documents Score, Noul, and Jev 1.14 Box bounding-box questions.
Documented context64k tokens total; state + longest question within 32k (262k opt-in)Default provider limits; TypeSafe-Danger-Zone-Long-Context: true enables 262k tokens. This public Playground applies smaller input and batch limits.
Official endpointPOST /v1/systemoneCall api.typesafe.ai from your server with a protected TypeSafe API key.
Pricing and capacity$42 / Btok ($0.042 / Mtok) input; output freeDocumented on the TypeSafe models page (2026-10-28) for Jev 1.14 and 1.13. Provider prices and rate limits can change; measure full workflow cost.
  • No speed, accuracy or cost advantage is claimed here. A confidence score describes the distribution for one request; it does not establish how often that answer is correct.

Checked against the TypeSafe model, API, Choice and confidence documentation listed in Sources.Verified

Capabilities

Decisions your application can inspect.

An explicit output space

Name the valid labels before a request. A Choice returns one of those labels, so your application can map it to an allowed route.

A distribution to inspect

Review how probability is distributed across choices. Test a fallback policy on labeled examples instead of interpreting a high score as proof.

A repeatable input contract

Send structured state and a question with label criteria. Keep task instructions separate from untrusted user text.

Application-owned actions

The model selects a label. Your code still validates permissions, checks policy and decides whether to execute a tool or request human review.

Try a decision

The same pattern, different tasks.

These examples load editable inputs into the Playground. Opening a sample does not send a model request.

Route a support ticket

A parcel is delayed and the customer wants a tracking update. Ask which team owns the request; provide the available support teams.

Try support routing
Illustration of incoming support messages classified into Billing, Shipping, and Technical teams, with a review route.

Choose the next interface control

A user needs a receipt. Compare the visible controls with that goal and select one allowed control, or No suitable control.

Try a browser decision
Illustration of context, one question, and a finite choice list feeding a decision model that returns a choice.

Check a claim against a source

A plan description excludes single sign-on. Decide whether a claim about access is supported, contradicted, or not settled by the supplied excerpt.

Try citation checking
Illustration of ranked choices with a highlighted leading result and a separate human review route.
Implementation path

Build a decision your code can inspect.

  1. Illustration of a delayed parcel and a customer message asking where the package is.

    Frame the context

    Include the information needed for this decision. Leave out unrelated logs and redact customer identifiers that do not affect the route.

  2. Illustration of one focused support-routing question: Which team?

    Ask one operational question

    Make the task specific: which team, which tool, or which evidence block? Avoid asking for an undefined best answer.

  3. Illustration of four allowed team choices: Billing, Refund, Shipping, and Technical.

    Define labels and criteria

    Describe when each allowed label applies. Add an unknown or review option for inputs that do not fit the ordinary routes.

  4. Illustrative Shipping choice with ranked result bars and a handoff to application code; not actual API output.

    Validate, evaluate, then act

    Check the response, calibrate a review policy on labeled examples, and measure the downstream outcome before enabling automatic actions.

Model FAQ

Jev model capabilities and limits.

What does the Jev decision model return?

The Choice workflow used by this site returns a selected label, probabilities for the supplied options and a confidence value. The broader TypeSafe API also documents Score, Noul, and Jev 1.14 Box question types. This Playground is a Choice experience; it does not generate a support reply or the answer to a routed task.

Is confidence the chance that the answer is correct?

No. TypeSafe documents confidence as the concentration of the answer distribution. A confident decision can still be wrong. Evaluate labeled examples and set your review policy using actual outcomes.

Can I upload a spreadsheet or image to the model?

This Playground accepts text and parses supported CSV, TSV, or TXT tables in the browser to send rows for classification; it does not accept image or XLSX uploads. In direct server calls to the TypeSafe API, Jev 1.14 and Jev 1.13 support opt-in image assets with the TypeSafe-Danger-Zone-Images: true header, and Jev 1.14 supports Box questions for image bounding boxes.

Should I use jev-latest in an evaluation?

It is convenient for trying the API, but the alias may change. Keep the resolved model identifier from each response along with your dataset and settings so results remain attributable.

Where do I start with an integration?

Try the Playground, then use the support-routing recipe for a server-side example. Keep the provider key off the browser and treat the returned route as a recommendation until your application has checked its own rules.

Test the decision before you automate it.

Bring a representative input, define the valid labels, and compare the result with the route you expect.