What is this site?
JevModel is a web console and API gateway for the Jev structured decision model. You send text or JSON state plus questions with your own options, and receive typed Choice, Score, or Noul answers with probabilities your app can branch on—without parsing free-form prose. The site offers limited no-signup live runs, free tokens after email verification, visual decision flows, batch evaluation, and an API on the same balance. It fits embedding classify / route / moderate steps into a product, not open-ended chat drafting.
Key features
- Browser playground with a few daily live runs without signup, then one-time free tokens after verification
- Three answer types: Choice among your labels, Score on a discrete scale, and Noul yes/no judgments with probabilities
- Decision Flows that chain questions on a canvas; one request runs the path and bills only reached branches
- Batch jobs on CSV / TXT / JSON Lines with templates or custom questions and exportable structured rows
- Shared API balance across Console, batch, and keys; failed requests are not charged or are refunded
- Freemium packs and subscriptions after the free tier; any purchase lifts free-account flow and batch-row caps
Common use cases
- Comment or community systems labeling toxicity, harassment, or spam before publish-or-block decisions
- Support inboxes tagging cancel, bug, or sales intent and routing tickets to the right queue
- Growth or sales scoring leads and deciding when a human should follow up
- Developers comparing question wordings on their own samples in batch before production thresholds
- Agent pipelines screening prompt-injection or personal-data signals before reading web or email content
Who is it for?
- Product and backend engineers who need labeled options plus probabilities, not long natural-language replies
- Ops or growth teams doing support triage, moderation, or lead scoring
- Indie developers validating question types in the browser before moving to the API
- Not a fit for users who mainly want long-form writing, brainstorming, or open chat
- Also not a fit if you only want a raw provider account and fully custom orchestration with no Console or batch (the site notes direct provider rates can win at very high single-call volume)
How does it compare?
Keep these distinctions in mind versus a general LLM workbench:
- Structured decisions vs free generation: Jev returns your options and probabilities; chat models lean toward prose that needs parsing
- Path-billed flows vs ask-every-question: Flows charge only reached nodes, which can cost less on deep trees in real traffic
- Integrated workbench vs bare API: Playground, Flow, Batch, and keys share one balance; for huge single-call volume, compare provider direct pricing too
- Trial friction: a few no-account runs plus verified free tokens suit testing schemas before paid scale
FAQ
Q: Can I try without signing up?
A: Yes. The homepage / playground allows a limited number of live runs per UTC day; verifying email grants a one-time free token grant (the site estimates coverage for short requests).
Q: How is the API billed?
A: It uses the same balance as the console, charging model-reported input tokens while output is free; failed runs without a saved result are refunded.
Q: Do tokens expire?
A: One-time packs do not expire; unused subscription tokens roll over, and yearly plans deliver the full year up front—see the pricing page.
Q: What limits apply on a free account?
A: Free accounts typically cap saved Decision Flows and batch row counts; any purchase lifts those caps as described on pricing.
