featherless-ai/simple-jev
Turn any open model into a classifier/jev endpoint
Laya and Jev are open models for typed decisions: instead of free-form text they return one choice out of a fixed set, fast enough to sit inside an agent loop. Laya is the local, on-device line with ports to Apple Core ML and other runtimes; Jev is TypeSafe AI's System One model served over an API, used for routing, review, browser control and MCP tools. This page tracks the GitHub projects, ports and integrations built on both, ranked by stars and refreshed with the rest of the boards every day.
Last updated: 2026-09-21 06:02 UTC · 12 open-source Laya and Jev projects, ranked by GitHub stars
Turn any open model into a classifier/jev endpoint
Build calibrated AI classifiers from human feedback using Jev and GEPA.
Local Laya typed decisions on Apple Core ML and Neural Engine. Validated ports, ~5 ms short decisions on M3 Max, reproducible speed and energy benchmarks.
Jev-powered model routing, memory, compaction, skill selection, computer and browser use for Hermes agents (also Claude Code and Codex)
Local-first MCP plugin for continuous software-quality review by AI coding agents, powered by Jev.
An educational Jev-like visual inference experiment on Apple Silicon: shared context, direct candidate scoring, and local visual demos.
Fast, cheap, typed judgments from TypeSafe's Jev model, as MCP tools.
Control a real browser by voice. Jev (TypeSafe System One) decides intent + target in ~300 ms per spoken word; Playwright acts — often before you finish the sentence.
Adapt local language models into Jev-compatible structured decision engines with Choice, Score, and Noul outputs powered by prefill-only binary inference.
An easy way to use jev with your coding agent for tool calling reasoning
Say it, and your Mac does it. A computer-use harness on Jev that reads the screen through Accessibility. Fast, no vision model
A tool calling chat bot built with Jev and no LLM.