Havuz

Community-driven benchmark suite for MLX inference engines on Apple Silicon

Open-source middleware for AI agents — catches mid-run failures,compresses stale context, and grows a reasoning library across runs. Any mod

MCP server that shrinks logs, JSON, docs and web pages before your AI reads them. Offline, deterministic, ~3ms, no API key. 66% fewer tokens

An open-source, self-hostable AI operating system and development workspace. Powered by Electron and React 19, it features multi-agent orche

Local-first Personal Memory OS on macOS: turn recordings, documents, and AI conversations into memory your agents can actually reuse.

Embedded long-term memory database for AI agents & LLM apps. Six-layer cognitive architecture, knowledge graph, hybrid BM25 + vector search,

A lightweight, zero-dependency Python checkpoint layer for structured LLM output reliability.

A self-organizing swarm of AI agents with no leader: shared task board, git workspace, agents that write their own MCP tools, and the swarm

GamePause for Local AI: a Windows tray app that pauses LM Studio, Ollama and other local AI servers while you game, and restores them afterw

Self-hosted agentic OS — a Vibe-Engineering platform for Builders. Connect Claude Code Agent to Discord, Telegram, WhatsApp, Slack & Email.

A simple studio to use agent

Local-first AI assistant for Windows: voice (Whisper+Piper), webcam vision, FLUX image gen, image-to-3D, memory, 80+ tools. OpenAI-compatibl

Let idle local LLMs sleep and get your Mac's memory back. A native macOS menu bar app to turn Ollama on and off and auto-free idle model RAM

The agent loop for Python. Your harness, any model.

Give your CMaNGOS/playerbots AI companions real personalities with a local LLM — a tiny, dependency-free Python shim over Ollama. Per-bot pe

Local-first agent harness (Windows + Ollama) for 2–9B models. Same 2B model: 0.017 → 0.821 across four agent harnesses — 48×. 288-cell bench

Local-model (Ollama) integration for DeepSeek Harness: discover, pull, remove, and inspect local models, route requests to them by task type

Your team uses Odoo — without logging in to Odoo. Free, self-hosted AI assistant on Telegram & WhatsApp: employees apply for leave, approve

Local-first YouTube research: acquire VTT, search with SQLite/FTS5, run human-gated cumulative research in a web app or CLI, query via read-

🛡️ Turnkey cybersecurity lab for Hermes Agent — 397 CVE-driven skills, 152 tools, 28 frameworks (MITRE ATT&CK, MISP, CVSS, EPSS, OWASP, NIST

AI assistant with persistent memory for Home Assistant. Voice & text via Assist, plain-language automations with confirm-first previews, one

Open-source local-first AI desktop copilot with an expressive blob companion, voice and text commands, browser automation, app control, and

The visual compute platform for AI and robotics — connecting agents, GPU workloads, simulation, training, and real robots through typed work

Privacy-first PII tokenization middleware for LLM pipelines — local detection, encrypted vault, zero cloud dependency.

A lossless token-optimization codec for LLM agents — cut input tokens up to 82%, nothing dropped, every fact kept. Drop-in proxy + one-line

Self-hosted AI gateway for small teams: a key per person, budgets that fall back to local models, PII masking and free SSO.

Pull 90-day news snapshots.

Local OCR, metadata automation and RAG document chat for Paperless-ngx using PaddleOCR and Ollama, designed for CPU-only self-hosted servers

Native AI workspace for macOS. Local and hosted models, agents, browser, terminal, voice, and mobile access. Locus is wallet-free; LocusX ad

Fully local AI assistant that watches your screen to eliminate the context problem. No cloud. No data leakage.

Local-first desktop coding agent with BYOK, MCP, and OpenAI-compatible / Ollama hosts

TYSA (Teach Yourself Anything): an learning harness that finds where your understanding ends and teaches from there. Electron app built on M

Open coding agent that matches the best at lower cost: any model, sandboxed, scriptable, embeddable. Engine of Murage.

Clónalo, ábrelo en Claude Code y un arquitecto de IA te entrevista, elige tu runtime y construye tu propio Jarvis: en tu compu o en Railway.

🔐 Replayable RAG — a local-first, auditable knowledge-reasoning platform. Graph-RAG + append-only audit log + deterministic contradiction ar

A self-hosted knowledge broker for homelabs. Routes and fuses queries across Kiwix, FreshRSS, Open-Meteo, SearXNG, Uptime Kuma, and Home Ass

🪝 透明可控的 AI Agent 终端 | 状态机 · 多层记忆系统 · 沙箱防御 · 完整审计 · 多人设多技能

Windows utility focused on ollama model manager, model library, and download local models for local AI workflows.

Windows utility focused on model download manager, resumable downloads, and local model queue for local AI workflows.

Self-hosted AI agent on K8s that reads live pods, events, and logs. Local LLM (Mistral/Ollama), FastAPI, read-only RBAC, ArgoCD + Image Upda

MCP server for Paperless-ngx — lets AI agents like Claude search, upload, tag, and organize your documents, with optional local semantic sea

One command, 28+ top AI CLIs — an interactive launcher for terminal AI coding agents (Claude Code, Codex, Gemini, Aider…). Uzbek &

Pure-Rust async runtime for LLM agents. Multi-provider (Anthropic/OpenAI/Ollama), tool-capable end-to-end, structured permissions, real MCP

Self-learning AI agent that gets smarter and cheaper over time. Routes between local and cloud LLMs, learns from every interaction, remember

Windows utility focused on local AI API server, local endpoint manager, and OpenAI compatible API for local AI workflows.

OpenAI-compatible proxy that keeps LLM agent context small: it injects relevant memories instead of the full chat history (−51% tokens on re

Windows utility focused on one-click local AI setup, local AI runtime, and Windows AI setup for local AI workflows.

Karpathy's LLM Wiki, built from your Claude Code sessions: a folder of markdown you own, Obsidian-native when you want it. Keeps what you de

Yumi the compagnon AI in your notch that asks before it acts. Reminders, calendar events, notes: shown exactly before, checked after. Nativ

Agent OS in Rust — one binary, 37 agents, 26 LLM providers, 37 channels (Telegram/Discord/Slack...), MCP server, A2A protocol, 120+ API endp