Jev
TypeSafe AI, Inc.
TypeSafe's "System One" model that returns typed judgments and probabilities instead of text, used via API.
Platforms
Overview
Jev is the first of what TypeSafe AI, Inc. calls "System One models", a new kind of AI model developed by the company. According to the official documentation, instead of generating text, it takes application state and a question and returns a typed answer with probabilities and confidence in the form of a Choice, a Score or a true/false judgment (Noul). The company cites uses such as routing, ranking, extraction and verifying LLM output, and states response times of 70 to 500 milliseconds. Published pricing is $0.042 per million input tokens with output free, and SDKs are available for Python and JavaScript. The official site announces that the waitlist was removed on September 27, 2026, so anyone can now create an account. It is not suited to chat or long-form text generation.
Main features
- Three question formats: Choice, Score and Noul
- Returns typed answers with probabilities and confidence
- Evaluates multiple questions in parallel in a single API call
- Python / JavaScript SDK
- Web-based Playground and console
- Agent skills for coding agents
Recommended use cases
Adding decision logic to apps
- Route inquiries by intent
- Re-rank search results
- Hierarchical classification
Checking LLM output
- Guardrails for LLM output
- Checking that citations are supported
- Classifying passages retrieved by RAG
Data processing
- Extracting dates and values
- Restoring the structure of malformed data
- Entity matching for knowledge graphs
Pricing
Pricing not yet verified — check the official website.
Screenshots
Related AI tools
Claude
Anthropic's conversational AI assistant, strong at long-document understanding and natural writing.
Gemini
Google's conversational AI, integrated with Search and Workspace.
Groq
An inference cloud built on its own LPU processor, operating dedicated capacity at scale.
OpenRouter
OpenRouter is an AI infrastructure and API service useful for AI development and model comparison.