Havuz

Open Source Instruction Tuning LLM Training Library 2026

Two source-verified AI-agent skills that turn project code into traceable business-rule documents (whole-project + single-module). Graph-gui

WorkBuddy Multi-Account Manager & Automated Check-in Toolkit (多账号轮换、全域工作区打通与自动化签到)

RF/microwave design automation + LLM orchestration: 16 EM/EDA engines behind one interface, 75 parameterized template families, 340 determin

Dohnuts builds small multimodal models for direct decisions. -> System One model

Myrtia — agent-native red teaming for LLMs and AI agents: a deliberating council of attacker models (propose, cross-critique, Borda vote, ch

Disaggregated quantization official codebase

Local, open-source Jev alternative: browser agent decisions with Laya (System One model) on your own machine. No cloud, no API key. Playwrig

Code for the paper "Emergent Collusion in Long-Horizon LLM Agent Interaction"

A Minesweeper benchmark for LLM agents. One identical board, up to nine models in parallel, one clock, one tool layer.

ZCode 桌面客户端增强插件：思考档位配置（3.14+ 原生 optionSpecs）、用量图表全量、TPS 状态栏、思考强度滑条、一键增强提示词、设置页模型拉取。纯 Python 标准库，全部可检查、可精确还原、可在插件详情页逐项开关。

网页版登录态 → OpenAI 兼容本地反代：DeepSeek / Qwen / Doubao 免费额度驱动 ZCode。Web login → OpenAI-compatible local API.

Craft Type-Safe Uncertainty-aware semantic pattern matching, control flow, and smart procedures for Effect DecisionModel and Jev

a DSL that blocks AI mistakes via proof

Point at any part of your running app, describe the change, and Lasso edits the real source file — not the DOM. Works with Vite and Next.js.

Measures how much code AI coding agents add, from patches benchmarks already publish

Practical guide to deepseek code: the Deep Code CLI for DeepSeek-V4 and DeepSeek Coder models on Ollama.

打开这个文件夹，就有一位一对一导师。A Markdown teaching protocol that turns any AI tool into a one-on-one tutor.

jevals, agent trace'lerini eval ve guardrail kontrollerinden tek istekte geçirerek production veya agent loop içinde değerlendirmeyi amaçlayan bir Python kütüphanesidir.

Neden yükseliyor · 7 günlük bir repoda 97 yıldız, 5 commit ve 4 farklı yazar erken ilgi ve katkı işareti. Tek istekte birden fazla kontrol, düşük maliyet ve agent loop'a uygun hız iddiaları dikkat çekiyor; performans ve maliyet rakamları README'de verilmiş, bağımsız doğrulama olarak kabul edilmemeli.

37 tok/s from Qwen3-Next-80B (48 GB) on a 32 GB laptop, 24 without drafts vs 2.8 for llama.cpp; reading 62 vs 16 tok/s. Faster than llama.cp

Jev-Mem: System-One Controlled Agentic Memory

A git hook that won't let you push code you can't explain. Quizzes you on your own diff. Works with Claude Code, Codex, Cursor.

安卓AI聊天端:开箱即聊，无损导入酒馆卡，缓存强省用量，手机操控，专属RP模式，防空回，主动信息，前端卡支持，高定制主题，QQbot，AI群聊，插件系统，强兼容中转站。手机移动端原生支持，求 Star 收藏⭐Android AI Chatbox: lossless SillyTa

Harness-Zero: Harness Distillation via Agent-as-Harness

Joy — a local-first personal assistant. Rust core (loop / memory / tools / graph / evals), TypeScript gateways + dashboard, Python thin SDK.

SpeakerMem-R1: Speaker-Centered Dual-Track Memory for Multi-Party Dialogue

TypeSafe Jev for DeepSeek Harness, the Model Context Protocol, and plain Node: typed judgments instead of prose, offline by default.

Evidence-backed knowledge graph construction with typed Jev relation decisions

Tiny LLMs, Q4, in the browser. WebGPU lab inspired by yangqi0/petitgpt.

Hemmingway-1: a 27B model that writes the way a person writes. Open weights, free for non-commercial use (CC BY-NC 4.0); commercial use by a

Local CLI and MCP server for compacting test and build output, with retrieval from a redacted local cache.

Jev-powered, rule-based grader for text files. Runs every rule against every line in parallel. No skimming, no missed lines.

SDK-first, independently maintained SemIf fork for fast, self-hosted semantic decisions.

Acta: A Decision-Centric Runtime for Agents

🚀 AI 全链路求职自动化平台：抓取→清洗→AI评估→简历改写→自动投递，四大招聘平台 + 飞书 ChatOps 指挥 | Industrial-grade AI job-hunting automation: scrape → clean → LLM eval → resume

The decision layer for your Rails app. A Rails-native wrapper around TypeSafe's Jev System One API: typed, calibrated decisions in your cont

Zusia, a study buddy inside Zotero: ask Claude Code, Codex or Antigravity about the paper you are reading. LaTeX, drawings, theorem links, g

Production-grade convergence & anti-overthinking steering extension for Pi Coding Agent (tuned for Bonsai 2 / Qwen)

Fast, typed Python framework for Jev-as-a-Judge/Decision-model evaluation for LLMs, RAG, and AI agents along with human-label calibration.

Record once, replay forever — a production-grade browser data-collection platform. A PageResearch Agent extension turns real interactions in

自托管的长篇写作台：Python 后端 + 网页版 + 安卓壳。多 Agent 分工写作、世界状态引擎、568 条 AI 味规则、一键听书。clone 下来就能用，稿子只存在你自己机器上。

Coding-agent tools on TypeSafe's Jev, measured in public. review-router: caught 13 of 13 CVE fixes a filename rule sent to a quick review. j

Jev judgment, not generation: prune long tool output, screen fetched pages for injected instructions, and gate completion claims inside Deep

High-throughput, robust native DuckDB extension for batched and streaming TypeSafe/Jev classification, scoring, and semantic predicates from

Ask your DuckDB tables questions in plain language. A DuckDB port of pg-jev, powered by TypeSafe's Jev.

Glyd, model ağırlıklarını ve KV cache'i bit düzeyinde kayıpsız biçimde sıkıştırıp GPU'da kullanmayı hedefleyen bir compression yazılımıdır.

Neden yükseliyor · README, yaygın bir maliyet sorununa doğrudan yanıt verdiğini ve 32B ile 72B modellerde daha az GPU ile çalışmaya dair ölçümler sunuyor; ancak 9 günlük repoda 170 yıldız ve 474 commit'in yalnızca 2 farklı yazardan gelmesi, ilginin ve geliştirme temposunun geniş bir katkıcı tabanına dayandığını göstermiyor. Yükselişin kesin kaynağı veya bağımsız benimsenme verisi bilinmiyor.

LLM2Jev, yerel language model’leri Jev uyumlu yapılandırılmış karar modelleri gibi kullanmaya yarayan bir Python aracıdır.

Neden yükseliyor · Sekiz günde 342 yıldız ve son 28 günde 25 commit dikkat çekiyor; README’deki JevBench ve gecikme sonuçları da ilgi çekebilir, ancak bu ölçümler proje tarafından raporlanmış ve bağımsız doğrulama bilgisi verilmemiş.

Run Laya, the open-source Jev-compatible System-1 decision model, from Node.js / TypeScript via ONNX Runtime

Loopcut: an AI agent that works inside Blender. A fork of Blender (GPL).

Agent/RAG/AI应用：Agent零基础入门学习笔记➕编程平台