convez.ai
MCP ServerAutomaçãoAvançado
Json55Hdz

OpenACM — Agente IA Autônomo Self-Hosted

Json55Hdz/OpenACM

Agente IA autônomo self-hosted que controla seu PC e integra com MCP servers

Ver no GitHub
BomPluginScore · de 0 a 100

Agente IA autônomo que roda localmente no seu PC, com capacidade de executar comandos, navegar a web, conectar a servidores MCP e gerenciar múltiplos canais de comunicação. Oferece suporte a sub-agentes, memória RAG persistente e um roteador local de intenções que otimiza custos de tokens. Totalmente privado e sem dependências em cloud.

Instalação

$npx convez-ai Json55Hdz/OpenACM

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
2
Forks
1
Acompanhando
2
Saúde
71%
Último commit
há 1 dia
Versão
v0.5.0
Issues abertas
0
Contribuidores
4

Casos de uso

  • Automação de tarefas repetitivas e fluxos de trabalho locais
  • Web scraping, navegação automatizada e testes de interface
  • Execução de código Python e scripts shell com estado persistente
  • Integração de múltiplas ferramentas via MCP servers sem sair do agente

Veredito

OpenACM é uma solução robusta e madura para quem precisa de automação inteligente sem depender de APIs cloud. Excelente para desenvolvedores que querem controle total sobre dados e ambiente, mas requer familiaridade com Python, Node.js e infraestrutura local.

README

A tradução para o português deste README ainda está na fila.

OpenACM — Open Autonomous Agent

OpenACM is a self-hosted autonomous AI agent that runs on your PC. It controls your local environment, writes and executes code, navigates the web, and connects to any MCP server — all through a modern web dashboard.

No subscriptions. No cloud dependency. Your data stays local.

Created and maintained by Jeison Hernandez / JsonProductions.
If you use or build on OpenACM, a credit or a star goes a long way.


What it can do

  • Run commands & code — executes shell commands and stateful Python (Jupyter kernel)
  • Browse the web — Playwright-powered browser automation: login, scrape, screenshot
  • MCP Server support — connect to any Model Context Protocol server (unity-mcp, filesystem, custom tools, etc.)
  • Multi-channel — chat via Web, Telegram, or Console, all sharing the same AI brain
  • Skills system — define reusable Markdown-based skills the AI triggers automatically
  • Sub-agents — spawn specialized agents that work in parallel on tasks
  • RAG memory — ChromaDB long-term memory that persists across conversations
  • Local intent router — hybrid local/cloud architecture that skips the LLM for simple commands (~5ms, no tokens spent)
  • Loop trace debugger — inspect every iteration: context size, tool calls, LLM timing, truncations

Quick Start

Option A — npm (recommended)

Requires Node.js 18+ and git.

npm i -g open-acm
openacm install   # clones the repo and runs full setup
openacm start     # launch OpenACM

Or without installing globally:

npx open-acm install
npx open-acm start

CLI commands: install · start · stop · status · update · repair · uninstall


Option B — One-liner

Linux / macOS:

curl -fsSL https://raw.githubusercontent.com/Json55Hdz/OpenACM/main/install.sh | bash

Windows (PowerShell as Administrator):

iwr -useb https://raw.githubusercontent.com/Json55Hdz/OpenACM/main/install.ps1 | iex

Option C — Manual clone

Prerequisites

  • Python 3.12+
  • Node.js 20+
  • An API key from any supported LLM provider

Windows

git clone https://github.com/Json55Hdz/OpenACM.git
cd OpenACM
.\setup.bat       # first time: installs everything and launches OpenACM

Next time just run:

.\run.bat

Linux / macOS

git clone https://github.com/Json55Hdz/OpenACM.git
cd OpenACM
chmod +x setup.sh run.sh
./setup.sh        # first time: installs everything and launches OpenACM

Next time just run:

./run.sh

Docker

docker-compose up -d --build
docker logs openacm   # your dashboard token is printed here

Open http://localhost:47821, paste the token, done.


First Launch

  1. The console prints your Dashboard Token — copy it
  2. Open http://localhost:47821
  3. Paste the token to log in
  4. Go to Configuration and add your LLM API key

Supported LLM Providers

Uses LiteLLM internally — any provider it supports works:

ProviderExample model
OpenAIgpt-4o, gpt-4o-mini
Anthropicclaude-opus-4-5, claude-sonnet-4-5
Google Geminigemini/gemini-2.0-flash
Groqgroq/llama-3.3-70b-versatile
Ollama (local)ollama/llama3.2
Any OpenAI-compatible APIconfigure custom base URL in settings

MCP Servers

Connect to any MCP server from the MCP Servers dashboard page:

ModeWhen to use
Remote HTTP (modern)unity-mcp, most modern servers — just paste the URL
Remote SSE (legacy)older SSE-based MCP servers
Local stdiorun a local process (npx @modelcontextprotocol/server-filesystem, etc.)

Once connected, the AI automatically sees and uses those tools.


Dashboard

PageWhat it does
DashboardReal-time stats, activity, live events
ChatMulti-channel conversations with tool call visibility
ToolsBrowse available tools and execution history
SkillsCreate and manage Markdown-based AI skills
AgentsManage sub-agents
MCP ServersConnect to external tool servers
TracesPer-request debugger: context size, tool timings, errors
ConfigurationLLM model, API keys, channels, preferences

Project Structure

OpenACM/
├── frontend/                   # React + Next.js dashboard
│   ├── app/                    # Page routes
│   ├── components/             # UI components
│   ├── hooks/                  # API and WebSocket hooks
│   └── stores/                 # Zustand state
├── src/openacm/
│   ├── core/
│   │   ├── brain.py            # Agentic loop + trace system
│   │   ├── llm_router.py       # LiteLLM interface + retries
│   │   ├── local_router.py     # Local intent classifier
│   │   ├── memory.py           # Conversation memory
│   │   └── rag.py              # Vector memory (ChromaDB)
│   ├── tools/
│   │   ├── mcp_client.py       # MCP server manager
│   │   ├── registry.py         # Tool registry
│   │   └── ...                 # Built-in tools
│   └── web/
│       └── server.py           # FastAPI server + WebSocket
├── skills/                     # Built-in skill definitions
├── config/                     # Local config (not committed)
├── setup.bat / setup.sh
└── run.bat / run.sh

Privacy

OpenACM is fully self-hosted. The only outbound traffic is what you explicitly trigger:

  • LLM API calls to the provider you configured
  • Telegram/Discord messages if you connect those channels
  • Browser requests when you ask it to visit a site

Everything else — conversations, API keys, files, memory — lives in data/ and config/ on your machine. Use Ollama for a fully offline setup.


System Requirements

MinimumRecommended
OSWindows 10 / Ubuntu 20.04 / macOS 12Windows 11 / Ubuntu 22.04
RAM8 GB16 GB
Storage5 GB10 GB
Python3.12+3.12+
Node.js20+20+

Contributing

Contributions are welcome. See CONTRIBUTING.md.


License

MIT — free to use, modify, and distribute.
Copyright (c) 2026 Jeison David Hernandez Pena (JsonProductions). All copies and derivatives must include the original copyright notice.

Ver o original no GitHub

Bom para

Relacionados