extremely-specific-council

A decision cache for TypeSafe Jev-class models — memoize decisions so repeats are free, deterministic, and shareable. One 2 MB binary.

Multi-axis writing quality checker powered by TypeSafe AI's Jev model. Separate named checks, each with its own verdict and confidence.

An input that becomes what you mean: one text box that morphs into the right UI as you type. Powered by TypeSafe Jev, works offline.

Recorded comparison of Jev, Terra, and Opus on 100 synthetic security-triage cases, five passes each, with a static inspectable dashboard

LLM-authored task-specific Jev harnesses with optional full-trajectory reward reflection and GEPA evolution.

Compile agent policy prose into deterministic verdict programs: narrow evidence questions for the model, the verdict computed in code. Install: npm i -g jev-compiler

Natural-language linter combining ast-grep selectors with Jev verdicts to catch misleading names, stale comments, and tests that miss their stated behavior, with fixture-calibrated cutoffs and diff-aware review.

Stops AI coding agents from claiming work is done without evidence. Deterministic hooks decide, TypeSafe's Jev advises. Append-only ledger, zero runtime dependencies.

Vercel's open agent framework, which ships Jev as the default evaluation model in its experimental evaluate path.