The fastest way to start
Open the TypeSafe API Console and sign in with Google or email. Jev no longer has a waitlist, and new users currently receive $5 in credit. In the official Playground, paste text as the state and add one question.
- Start with Noul to detect whether one property is present.
- Then use Choice for classification and Score for an ordered rating.
- Test a few atomic questions first, then expand.
Understand one request
The state is the context Jev evaluates; questions define the permitted outputs; model usually uses jev-latest. Jev ingests the state once and evaluates all questions in parallel.
curl -X POST https://api.typesafe.ai/v1/systemone \
-H "Authorization: Bearer $TYPESAFE_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"state": "Checkout has failed three times and we are losing sales.",
"model": "jev-latest",
"questions": {
"urgent": {
"type": "noul",
"instructions": "The message is urgent"
}
}
}'Design questions Jev can judge
- Split “What is this customer like?” into urgency, sentiment, topic, and routing.
- Describe boundaries in criteria instead of relying on vague option names.
- Use confidence thresholds to choose automation or human review.
- Revalidate thresholds after upgrades; pin a version in production when needed.
Before production
- Build an evaluation set from real historical data.
- Log model version, input, output, and the eventual business outcome.
- Evaluate Chinese workloads separately instead of assuming English performance.
- Create fallback paths for timeouts, rate limits, and low confidence.
Never expose your API key in browser code. Call the TypeSafe API from your own server.
Continue reading
Explore real use cases
See exactly which decision Jev owns inside a larger workflow.