Jev 101 is an independent beginner-friendly guide, resource library, and interactive lab for typed AI decisions.

• Learn Jev: Start with comprehensive beginner guides that explain the core concepts of state, questions, choices, scores, noul, and confidence.
• Resource Library: Browse a verified library of official documentation, SDKs, tools, and community projects to deepen your understanding.
• Interactive Playground: Evaluate live decisions in a workbench environment. Adjust state payloads, tweak evaluation criteria, and inspect real calibrated outputs.
• Schema Builder: Understand and build question schemas for constrained decision targets.
• Practical Recipes: Discover blueprints that turn use cases into runnable examples, such as classification, security gating, and lead scoring.
• Transparent Benchmarks: Review small, transparent benchmark runs with exposed dataset versions, provider models, latency, and failure cases.
• Typed vs. Generative AI: Understand why System One models outperform LLMs for software decisions by providing typed values and calibrated probabilities, ensuring 100% format reliability and significantly lower inference latency compared to traditional LLMs.
• Key Features: Features include a continuous state transition loop, immutable typed data, calibrated probability outputs with type-checked certainty, and a clear decision anatomy with 3 primitives (State, Question, Verified Output).