ИИ-агенты
octomind
An MCP-native AI coding agent and runtime in a single binary.
★ 151 звездаApache-2.0
Смотреть руководство →arcbox — проект с открытым исходным кодом из категории «ИИ-агенты» от arcboxlabs (arcboxlabs/arcbox на GitHub). У него 4 226 звёзд на GitHub и 188 форков, он распространяется по лицензии Apache-2.0 и последний раз обновлялся на GitHub 7 октября 2026.
ArcBox: a fast macOS-native container and VM runtime — a drop-in Docker engine with agent sandboxes
Confirm macOS on Apple silicon (M3+ and macOS 15+ if you want agent sandboxes).
Install via the Homebrew cask or the curl installer.
Start abctl's daemon and enable the Docker engine so your existing docker commands work.
Run a normal docker container, or spin up a disposable agent sandbox with checkpoint/restore.
docker, Compose, BuildKit, volumes, and port forwarding work unchanged; amd64 images run on Apple silicon via FEX translation.
A built-in k3s cluster you can start and enable from abctl.
Disposable sandboxes (abctl claude, sandbox create/run) with checkpoint and restore for AI agents.
Create and SSH into Linux machines, and run throwaway macOS guests (up to Apple's two-guest limit).
Import your setup from Docker Desktop or OrbStack with abctl migrate.
No account and no cloud required; everything runs on your Mac under the abctl CLI.
Clone the project and enter its directory (most users install the binary, next step).
git clone https://github.com/arcboxlabs/arcbox.git
cd arcboxInstall the macOS runtime via Homebrew (or the curl installer from get.arcbox.dev).
brew install --cask arcboxlabs/tap/arcboxStart the daemon, enable Docker compatibility, and run a container as usual.
abctl daemon start
abctl docker enable
docker run -d -p 8080:80 nginxarcboxlabs develops ArcBox (arcbox.dev); it is currently in public beta.
The project is distributed via Homebrew/installer as a public beta; the repository metadata lists the Apache-2.0 license.
No — the README emphasizes no account or cloud connection is required.
It aims to be a drop-in replacement for both on macOS, adding native k3s, full VMs, and agent microVM sandboxes with checkpoint/restore, with migration commands from either tool.
Run `brew install --cask arcboxlabs/tap/arcbox`, then `abctl daemon start` and `abctl docker enable`, and use docker normally.
An MCP-native AI coding agent and runtime in a single binary.
Osaurus — own your AI: a native macOS agent harness with any model, memory and fully offline execution.
An open-source multimodal retrieval engine for agent pipelines.
A lightweight, container-isolated personal AI agent for WhatsApp, Telegram, Slack and more
Local-first microVMs for fast, programmable sandboxes.
Offline multi-agent simulation and prediction engine: Neo4j plus Ollama, local.
Другие проекты из категории «ИИ-агенты» с наибольшим числом звёзд в каталоге LibreAI — реалистичный список альтернатив, если arcbox вам не подходит.
A personal AI that really does things — on any OS, with your data staying yours.
An agentic skills framework & software development methodology that works.
Agent harness optimization system with skills, memory and security for Claude Code, Codex and more
The agent that grows with you — a Python framework for adaptive AI agents.
DeepSeek's plugin-based agent harness where everything is a plugin.
Соседние проекты из категории «ИИ-агенты» в каталоге, ближайшие по числу звёзд к arcbox — продолжайте изучать полку по обе стороны от него.
Monid — OpenRouter for agent tools: one base URL and one key give an agent 2,000+ tools across 72+ providers
AdalFlow: The library to build & auto-optimize LLM applications.
Archestra: the enterprise AI platform — chat UI, LLM gateway, and MCP gateway with guardrails
Excel MCP Server: let AI agents read, write, format, and chart real .xlsx workbooks — no Excel required
nono: zero-trust, zero-setup, zero-latency sandboxing and runtime security for AI agents
Open-source, vision-first browser agent
Бесплатно и открыто — забирайте исходный код, ставьте звезду репозиторию и начинайте создавать.