jev-model

Turn any LLM into a Jev-style decision model: typed decisions, real probabilities, no training. (continue updating)

A tiny jev-like model that answers Choice, Score and Noul questions in one forward pass and returns calibrated probabilities. MLX or PyTorch, fully offline, System One compatible.

Calibration-aware reinforcement learning for adaptive decision systems

Teach your agent to work with evals: WHEN you actually need an eval or benchmark, HOW to build one that holds up, and how to read what it tells you. Deterministic-first, tool-agnostic.

Natural-language constraints for JEPA world-model planning, judged by a decision model instead of an LLM.

A simplified functionality reproduction of Jev-style decisions with Gemma 4. Native multi-modal ability, finetuning free, runs locally on your machine.

Code and data for evaluating Jev, a System One model, on scientific decisions and how its choices affect downstream results.

Windows port of Laya typed-decision AI (ONNX Runtime + DirectML) — Core ML/Apple Neural Engine alternative with first-class Arabic support. No text generation, no hallucination, runs on any DX12 GPU.

an Information-Geometric Analysis Toolkit for any System-one (Jev, Jevlike) agent systems

An open-source, small and fast Jev-style model for structured text and vision decisions.

No yap, only fax. bruv a CLI tool that runs Jev, Simple Jev and other supported open models to turn your data into a clear yes, choice, or score.

decision-first data cleaning system powered by Jev

LISA: Leak, Injection & Simplicity Auditor. GitHub Action that uses TypeSafe Jev (a system one model) to flag secrets, security vulnerabilities, and unneeded complexity in pull requests

When Jev meets LLM — Transparent proxy that makes AI coding agents 3× faster and 80% cheaper. Cut response times from 2.4s to 890ms with Line J architecture.

Security requirements analysis with JEV, LangChain Deep Agent review via LLM, and deterministic Python reporting

Clanker frustration insights, powered by on-device Laya decision model.

Built to make fast, structured decisions that software can use directly. Built on top of Qwen3 you can train and run on your own.

Dockerized REST service, jev-compatible, that serves the Laya model (convaiinnovations/laya) for use from non-Python apps.

JevKit is an open-source decision infrastructure that helps developers integrate Jev into their applications, define structured decision tasks, evaluate model performance, and build reliable fallback policies.

Jev Guard is a real-time Windows process monitoring application powered by TypeSafe's Jev AI model. It tracks newly launched processes, logs event metadata (paths, command lines, and parent processes) to a local SQLite database, and classifies risk levels—labeling events as benign, suspicious, or malicious on a live interactive dashboard.