Jev AI Model
Jev AI Model turns text or JSON into structured choices, scores, and yes-no decisions with probabilities
About Jev AI Model
Jev AI Model is an AI decision workbench and paid API for developers. It helps classify, score, and route text by letting users define a state, ask typed questions, and inspect structured answers with probabilities.
The browser playground is free after sign-in and is designed for testing decision workflows before connecting an application. The homepage shows example uses such as routing customer requests, choosing an agent’s next step, and checking a claim against supplied evidence.
The API uses POST /v1/systemone with bearer-token authentication. Documentation says requests accept a model, shared state, and 1-8 named questions, with supported question types including choice, score, and noul yes-no checks.
Pricing is credit-based for API use. Web runs are free after sign-in, while API credit packs include Starter at $10 for 100,000 credits, Pro at $100 for 1,000,000 credits, and Volume at $1,000 for 11,000,000 credits
Features
- Classify and score text
- Define typed questions for a scenario
- Inspect structured decisions with probabilities
- API requests are limited to 32 KiB for the entire request
- Score questions support 2-10 ordered criteria levels
Use cases
- Classify incoming text for an application workflow
- Score text against scenario-specific questions
- Route text based on structured decisions
- Compare two decision configurations against up to 25 labeled cases
- - Move tested workbench prompts into a backend service through the documented API
Screenshots
Frequently asked questions
Can I try Jev AI Model for free?
Yes. The online playground is free after you sign in. API use is pay-as-you-go with credits.
How do I connect Jev AI Model to my software?
Connect through the Jev AI Model API and pay for API use with credits.
What does Jev AI Model return?
A structured decision with probabilities based on the typed questions you define.
Are failed provider calls billed?
No. Documentation says failed provider calls are not billed.
What history is stored?
Documentation says saved configurations and web request bodies/results are private, while API history stores metadata, token usage, and credits without request content.
0 comments
Sign in to leave a comment.
No comments yet. Say something useful and be the first.