Jev directory
CommunityJev Tools & CLIs
OpenRouter: typesafe/jev-1.13
Access Jev through OpenRouter's decisions endpoint, an alternative route to early access that still uses TypeSafe's question and answer shapes.
- Reading time
- 2 min
- Published by
- Community
- Last verified
- Sep 2026
Works forJev for Routing
Used inJev for Software & DevTools
Highlights
- Model ID typesafe/jev-1.13, with ~typesafe/jev-latest tracking the newest release; no waitlist or TypeSafe account is needed.
- Decisions API posts to /api/alpha/decisions; /api/v1/systemone accepts the TypeSafe SDK pointed at OpenRouter.
- Context window is 32,000 tokens for state plus questions; output tokens are free and usage.cost reports each call's USD cost.
- The sample response returned model typesafe/jev-1.13-20260917 with a noul of 0.96 and a usage cost of 0.000019992.
- All questions in a request are answered in parallel and cannot see each other's answers; Jev returns no reasoning trace.
Quickstart
bash
curl https://openrouter.ai/api/alpha/decisions \
-H "Authorization: Bearer $OPENROUTER_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "typesafe/jev-1.13",
"state": { "customer_tier": "enterprise", "ticket": "My checkout page shows a blank screen after I click Pay. I have tried two browsers." },
"questions": {
"is_bug": {
"type": "noul",
"instructions": "Is the customer reporting a software defect?"
}
}
}'Watch out
Calls are billed to your OpenRouter account at the input price on the model page, and confidence describes the distribution of alternatives rather than whether an action is safe.
Questions about this entry
Short answers drawn from the linked project or page.
No. OpenRouter is an alternative route that needs only an OpenRouter key, using model ID typesafe/jev-1.13, with ~typesafe/jev-latest tracking the newest release. Billing still follows the input price on the model page and lands on your OpenRouter bill.
It runs on the OpenRouter Decisions API, not the OpenAI-compatible chat endpoint, so request and response shapes are specific to structured decisions. POST to /api/alpha/decisions, or point the TypeSafe SDK at /api/v1/systemone with the same question and answer shapes.
Nothing to fail over to. The model is hosted by a single provider and OpenRouter forwards every request straight to it, so there is no routing decision and no second provider. The card reports 100.00% uptime over three days and 99.99% availability for that one provider.
No reasoning trace is returned. You get the answer with a probability for every option, all questions answered in parallel, plus usage.cost per call. The sample response on the page shows a noul of 0.96 at a usage cost of 0.000019992 on a 32,000-token context window.
The model page lists a 64K context and $0.042 per 1M input with $0 per 1M output. A model note says 32,000 tokens are available for state plus questions, and output tokens are free.
Every call reports a USD cost in usage.cost. The sample response returned model typesafe/jev-1.13-20260917 with a noul of 0.96 and a usage cost of 0.000019992.
The model ID is typesafe/jev-1.13, and an alias such as typesafe/jev-latest tracks the newest release. Neither requires a waitlist or a TypeSafe account.
From the Jev guides
Original write-ups on what this entry is for.
From the Jev AI model community
Posts from builders shipping with Jev right now.