Jev is TypeSafe’s model for focused decisions. Try classifying a message, compare the allowed outcomes, and see where human review belongs.
1 anonymous attempt per day · 20 signup credits · No card for the trial
Independent workspace powered by Jev through OpenRouter. Not affiliated with TypeSafe.
One label.
Confidence you can inspect.
A review path when it matters.
Illustrative sample — not a live model response.
Run a decision or open the labeled sample to see the result here.
A high score is not a guarantee. Keep a person in control of consequential actions.
Check this result, then test a small batch of your own messages. Each completed row uses one credit. Compare plans when you need API access; a credit pack is for web tools only.
A categorised index of community submissions: agents, developer tools, interfaces, games and experiments. Each entry is one line about what it does and a link to the maker’s own post. This is an independent workspace, not the vendor.
Read the guide ↗Turn an incoming customer message into one of the queues your team already uses. Keep the category list small, the routing rule explicit, and the final action under your control.
Read the guide ↗A good qualification rule explains why a lead belongs in a queue. Evaluate the information a prospect actually provided against a short, written definition of fit.
Read the guide ↗Choose the outcomes your team actually uses.
See the label, confidence when available, and review flag.
Use the API when the result is ready for your process.
Explicit labels, visible failures, and idempotency keys. Your application stays in control of every action.
curl --fail-with-body --connect-timeout 5 --max-time 35 \
https://jevapi.pro/api/v1/decisions \
-H 'Authorization: Bearer YOUR_API_KEY' \
-H 'Content-Type: application/json' \
-H 'Idempotency-Key: ticket-123-v1' \
-d '{
"text": "I was charged twice this month.",
"instructions": "Choose the responsible team.",
"labels": ["billing", "technical", "account"],
"threshold": 0.85
}'Integration example. Replace YOUR_API_KEY with a key stored on your server.
Start with the output your application needs. A support queue needs a category; a customer reply needs generated language. Jev and a generative LLM address those different jobs. This comparison helps you design and evaluate that boundary. Jev API Pro is an independent workspace that accesses Jev through OpenRouter; it is not TypeSafe's official product.
Read the guide ↗System One is TypeSafe's name for a model interface centered on bounded decisions rather than generated prose. Jev is its published model family. The useful question for a developer is concrete: what state goes in, what typed answer comes out, and who controls the next action? This guide explains those boundaries without treating the name as an industry standard or a guarantee of human-like judgment.
Read the guide ↗TypeSafe develops Jev, a model for typed decisions. Jev API Pro is an independent workspace using Jev through OpenRouter; it is not TypeSafe's official website or billing console. This guide separates the model provider, access routes, and workspace features so you can choose an integration and understand what you are paying for. Sources were checked on 25 September 2026.
Read the guide ↗No. Jev API Pro is an independent workspace. Jev is a model from TypeSafe, accessed here through OpenRouter.
Yes, when live decisions are available: one trial attempt per source address per UTC day, subject to shared trial capacity. A labeled example is also available.
We show an explicit error. Failed account decisions are refunded, and we do not substitute a different model silently.
Decision history stores the result, request hash, and operational metadata, not the submitted message. Your text is sent to OpenRouter and its model provider for inference.
One completed classification within the published input limits uses one credit. Signed-in accounts receive 20 signup credits and can claim two daily check-in credits.
Use the API to get decisions. Your own application checks permissions and performs actions. Low-confidence or missing-confidence results should go to review.
Answers come from the published product guide. For account-specific questions, contact support.
Contact