Havuz

Open-Source Agentic Harness for Reliable Task Completion, even on small local models. 💻🧠🤖

The Open Source Strix Halo AI engine. Qwen 3.8 Flash Next Q4_K_XL: 1,801.01 pp, 73.17 tg single user, 194.37 tok/s 8 users; Qwen 3.8 27B Q4

⚒️ Open-source game wiki template. Astro + Cloudflare Pages. Free, fast, beginner-friendly. Static-first, zero-JS, AI-native content workflo

Open-source AI-assisted HR onboarding for Feishu/Lark: configurable workflows, document OCR and review, Bitable sync, reminders, and a zero-

一个面向 AI Agent 与 LLM 集成的轻量级 Java 框架，提供统一抽象、OpenAI 兼容对话支持，以及流式响应、工具调用、记忆窗口和 Agent 编排的模块化结构。

更智能的车万女仆：排班调度 / 蓝图建造 / 挖矿伐木烧制酿造 / 战斗自保与单兵战术 / LLM 对话·长期记忆·17 个 AI 工具 / TTS 语音（1.20.1 Forge & 1.21.1 NeoForge；纯 TLM 前置）

An automated, multi-turn LLM red-teaming pipeline that pits an Attacker model against a Target model to evaluate safety boundaries, featurin

A tap-to-explore picture book where an AI draws every page in real time — type anything, click anything inside, and it draws a new page abou

Open-source agentic software development platform for building, orchestrating and monitoring autonomous AI coding agents.

AI agent workflow platform — visual flow orchestration, drag-and-drop forms, knowledge base RAG, multi-model LLM, digital workers, Tauri des

The AI agent with a memory that lives on your machine and answers the phone.

SGME 拾光记忆引擎：AI 全面接管记忆、技能库、WIKI 知识库三大模块，跨会话、多智能体共享记忆，AI 会一直记得你的偏好。ShiGuang Memory Engine · AI takes full charge of three core modules — memor

Open-source, local-first alternative to NotebookLM. Chat with PDFs, links, YouTube videos and Arabic articles; every answer cites the exact

A swarm of AI users that test your website in real browsers — visualized as a pixel-art aquarium. Live friction maps, multi-user scenes, up

A tiny C runtime for durable AI agents: local-first, event-driven, restart-safe, and fully inspectable.

AI 面试助手: 素材整理 / 全程记录 / 自动复盘 / 模拟练习。让每一场面试都变成下一场的准备。

OAK, C++ ile cihaz üzerinde AI agent çalıştırmayı hedefleyen, bazı agent özelliklerini açık kaynak modüller olarak sunan alpha aşamasındaki bir projedir.

Neden yükseliyor · 48 günde 1170 yıldız almış; on-device AI, gizlilik ve Qualcomm NPU desteği ilgi çekici bir konumlandırma. Ancak yükselişin özel nedenini mevcut veriden belirlemek mümkün değil. Son 28 günde commit ve farklı yazar sayısı sıfır; README’deki performans ve karşılaştırma iddiaları da burada bağımsız doğrulanmış kanıt sunmuyor.

Ephemeral Workspaces, durable shared history. LayerFS gives every agent a complete, writable filesystem Workspace over shared committed hist

智能微信聊天机器人（基于 KouriChat 修改，兼容微信 4.1.12+，底层使用 wechatauto-replica） - WeChatBot_WXAUTO_SE

Covert Coder — a local-first sovereign AI development environment for governed models, workflows, execution, evidence, and verification.

Research skill and researcher subagent for Claude Code: multi-engine search behind one interface, parallel agent waves, and a verify gate th

Local-first MCP memory server for AI agents. Zero dependencies, SQLite on your own disk, 52 MCP tools with Mem0-compatible names. MIT.

Open-source vector database and sub-microsecond key-value store in one Go engine — embed it as a library, run it standalone, or replicate it

Source-linked readings and reproducible studies of agent memory, belief revision, and experience reuse.

Open-source cost metering for n8n AI workflows — token usage per workflow, priced daily, on your own machine.

A self-running research agent that monitors the web on your schedule, verifies findings with LLM analysis, and delivers intelligence reports

DeepSeek-V4-Flash inference for AMD Strix Halo

Self-hostable reading-position sync and statistics server for Liseur, KOReader, and friends

User Interface for Prime-Agent

True incremental Markdown-to-ANSI terminal renderer for streaming LLM output. CommonMark/GFM support tested, no full-document buffering. C C

Real-time Windows system-audio translation captions for Chinese, Japanese, English, and Korean, with optional LLM revision.

Навык для ИИ-агентов (Claude Code, Codex, Cursor): упрощает русский технический текст по ГОСТ Р 58049—2017 — комментарии в коде, документаци

The Brain of Theseus — a plain-text roguelike written to be played by an AI, over MCP (stdio). Reincarnation, memory, and irreversible cyber

Self-hosted AI presentation generator: topic or sources in, researched themed editable PPTX and report out. Docker, BYOK or Ollama; 74 slide

Statewave use case in TypeScript: a grounded product advisor and support assistant, with an Ops Assistant that tracks and closes content gap

The document extraction benchmark by Flexprice

HarnessRouter, farklı agent harness’lerini oturum, streaming ve dosya desteğiyle tek bir API üzerinden kullanmaya yarayan self-hosted bir altyapıdır.

Neden yükseliyor · Agent harness’lerini tek arayüzden kullanma vaadi güncel bir altyapı ihtiyacına hitap ediyor. 49 günlük repo için 2.675 yıldız ve son 28 günde 271 commit ile 10 farklı yazar, belirgin ilgi ve kayda değer geliştirme hareketi gösteriyor; ancak bu metrikler benimsemenin veya üretim kalitesinin kanıtı değil.

unlazy, AI agent’ların işi tamamlandı demeden önce kabul ölçütlerini yazmasını, kontrolleri çalıştırmasını ve sonuçları yeniden doğrulamasını sağlayan bir skill ve Node.js denetleyicisi sunuyor.

Neden yükseliyor · 49 günde 3709 yıldız dikkat çekici bir ilgiye işaret ediyor; ancak bu ilginin nedenini verilen verilerden çıkarmak mümkün değil. Son 28 günde 1 commit ve 1 farklı yazar olması, yıldız artışının güncel geliştirme temposuyla desteklendiğine dair kanıtı sınırlıyor.

Your car as a chat-room agent: Raspberry Pi 5 + dashcam + local AI. CodeWatch's sibling for the garage.

Local-first YOLO workbench for annotation, training, evaluation, and AI assistance.

Drive your real Chrome from any AI agent — logged-in sessions, real fingerprint, live tabs. CLI-based, works with OpenCode, Pi, or any Harne

A workbench for language model research

自托管 AI Coding Agent（Rust）：常驻 Runtime 撑起终端 / 浏览器 / 手机三端；每行改动可追溯到具体 Agent 与工具调用，子 Agent 的活由退出码裁决，检查点回退、目标 token 预算与验收清单跨轮次跨重启累计，用量与反馈账本留在本地。接任

Mako — AI Career Intelligence System

A stock picking tool that has 21 legendary investors like Warren Buffet, Peter Lynch, Michael Burry etc debate every stock pick and show you

Voice input that transcribes what you said, grounded in what's on your screen. macOS, Windows, Android, iOS.

A cache-aware attention and forgetting layer for high-frequency everyday AI chat.

Small, independent TypeScript packages for LLM plumbing — token budgets, streaming JSON repair, cost accounting, retries, embedding caches.

Open-source infrastructure for the full inference lifecycle: deploy, observe, scale, optimize, and safely release self-hosted models behind

Self-hosted AI voice agent for Asterisk. Streaming STT to LLM to TTS over AudioSocket, with barge-in and tool calling.