Ray-Hughes/jevalyn

Semantic decisions as ordinary Ruby #feels?, #decide, #score, Rails validations and pattern matching powered by Jev

Unofficial TypeSafe Jev showcase — System One decisions, not chat.

Ruby client for decision models such as Typesafe Jev

Probabilistic conditionals for Ruby: Feelings(message).like?("genuinely urgent")

Jev-like family of decision models built on top of Qwen3.5/3.8 you can train and run on your own

The open, local take on Jev: typed decisions from an LLM, without generating a single token

Open, calibrated System 1 decision models: typed choice / score / yes-no answers with a probability for every option. 400M to 80B; CPU, NVIDIA, Apple Silicon, LM Studio, Ollama and vLLM.

Open replica of TypeSafe's Jev: typed calibrated decisions in one forward pass, on Gemma 4 E2B / Gemma 3 270M (Modal)

OpenSource Jev