Playground/Language & data/Example builder
API ACTIVITY

Usage & budget

PLAYGROUND SAFETY LIMITS

Steady requests. Visible limits.

0 active · 0/20 queued

Browser: one start every 1.2 seconds (at most 50/minute), up to 2 active calls and 20 waiting. Shared by examples in this tab. Waiting requests cancel when their run is stopped; a full queue rejects new work without an API call.

Server: 60 requests per rolling 60 seconds and 2 concurrent calls per key, per server instance. Other tabs using the same key share that instance’s allowance. This in-memory safeguard resets on restart and is not a distributed or account-wide quota.

Source: playground policy in lib/requestRateLimit.ts. TypeSafe may impose separate limits. Provider HTTP 429 pauses calls until its reported reset; without a reset time, retry requires your action. HTTP 402 is a budget/billing failure, not a pacing limit. No automatic retries or background replay.

Your TypeSafe API key

Import your key to override the shared server default.

Saved on this browser until removed or site data is cleared. Masked on screen; sent only with Jev requests through this app. The saved value is never displayed.

Connecting

Test setup

WORKSPACE GUIDE

Example builder

Explore the original catalog, edit inputs and define your own closed-set questions.

ContextTyped questionsAnswers

What you provide

An example's Text or JSON context and selected typed questions. A/B cases declare the single field that changes.

What happens

The request sends your context with noul, choice, or score contracts. Jev returns structured answers for those questions; drafts stay in your browser.

What you get

Answer cards with model confidence, comparisons, and reference notes where provided. Confidence and a typed response do not establish correctness.

Walk through a run

  1. 01

    Pick a question

    Browse a collection or search the 110 starter examples. Your edits stay in this browser as drafts.

  2. 02

    Change the context

    Edit the input and choose which questions to include. For A/B examples, inspect the one field that changes.

  3. 03

    Read the evidence

    Run the example, compare the typed answers, and open reference notes where available. Export keeps your API key separate.

Try this

Open an A/B example, inspect the changed field, then compare both answers. Reveal reference notes after making your own prediction.

Execution limits

Runs use live Jev. Confidence is a model signal, not proof of correctness. Subjective examples have no universal answer key.

Explore the example. Check the evidence.
Original example
Use cases / Social & community

A real question in the replies

Separate genuine interest from generic engagement bait.

01

Provide the context

Edit the example’s input to test your own scenario.

02

Choose what Jev evaluates

3 of 3 questions included. Expand a question to edit its instructions.

Reply intentChoose one
{
question:"Genuine request for information"
correction:"Relevant factual correction"
joke:"Humor or playful sarcasm"
promotion:"Unsolicited promotion or spam"
hostility:"An attack meant to provoke"
other:"None of the above"
}
Worth answeringProbability
TensionScore
[
0:"Constructive or neutral"
1:"Mildly pointed"
2:"Personal attack"
3:"Threatening or intensely hostile"
]
Edit all questions as JSON

Edit types, candidate definitions, or add questions. Apply before running.

Try this

Replace the reply with a generic compliment and compare.

3 questions · Run: 1 request · jev-latest

Results

ContextTyped questionsAnswers

Ready to experiment.

Choose an example, inspect the input, and run it to see typed results here.