AmberClaw AI OS
krishujeniya/AmberClaw
Sistema operacional IA com orquestração kernel, governança zero-trust e suporte edge
Ver no GitHubFramework de assistente IA pessoal leve e modular que funciona como um sistema operacional para processos IA contínuos. Suporta múltiplos LLMs (OpenAI, Anthropic, Gemini, Ollama local), execução em sandbox, memória persistente em 3 camadas e integração nativa com hardware (Raspberry Pi, Arduino, ESP32).
Instalação
npx convez-ai krishujeniya/AmberClaw1 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
Casos de uso
- Assistentes IA autônomos que rodam continuamente em background
- Integração de IA com IoT e dispositivos embarcados (RPi, Arduino)
- Aplicações que precisam rastreamento de contexto longo e memória persistente
- Deployments seguros com aprovação humana e sandboxing de código
Veredito

Projeto robusto e bem-arquitetado para quem quer orquestrar agentes IA em produção. A stack é pesada (asyncio, FastAPI, Kubernetes) mas as abstrações justificam — zero-trust governance e suporte hardware são diferenciais reais. Recomendado para builders que precisam de controle total e segurança.
README
A tradução para o português deste README ainda está na fila.
AmberClaw AI OS
AmberClaw AI OS is a fully modernized, enterprise-grade personal AI assistant framework designed for 2026 standards. It acts as an operating system that runs continuous background AI processes, tracks long-term context using a 3-layer memory architecture, and ensures absolute safety with zero-trust execution environments.
Features
- Kernel-Level Orchestration: Fast asynchronous
HeartbeatEngineschedules and tracks subagent lifecycles natively via Pythonasyncio. - Zero-Trust Governance: The
GovernanceBoardhalts risky or expensive operations for human-in-the-loop approval. - Sandboxed Execution: Subagents write and execute code in isolated containers via
CodeSandbox. - Ubiquitous Memory: 3-layer semantic persistence powered by
mem0aiandChromaDB. - Model Agnostic Routing: Utilize OpenAI, Anthropic, Gemini, or Local Ollama models interchangeably via
LiteLLM. - ClawHardware Integration: Native support for Raspberry Pi GPIO, and serial communication with Arduino/ESP32 peripherals.
Hardware & Edge Support
AmberClaw is designed to run as the primary AI OS on edge devices.
🍓 Raspberry Pi (ClawOS)
To turn your RPi into an autonomous AI node:
./scripts/setup_rpi.sh
This script installs dependencies, optimizes the Python environment for ARM, and sets up AmberClaw as a systemd service.
🔌 Microcontrollers (Arduino/ESP32)
AmberClaw can control peripheral hardware nodes via Serial.
- Flash your device with the provided firmware:
examples/hardware/amber_node.ino. - Connect via USB.
- Use the
hardware_send_commandorhardware_read_sensortools to interact with your physical environment.
Quickstart
1. Installation
We use uv for lightning-fast dependency management.
# Clone the repository
git clone https://github.com/krishujeniya/AmberClaw.git
cd AmberClaw
# Copy env template
cp .env.example .env
# Install dependencies and sync virtual environment
make install
2. Booting the OS
Start the AmberClaw Gateway and core OS engines:
make dev
Alternatively, check the system status via the Typer CLI:
make cli
Production Deployment
We ship production-ready configurations out of the box.
Docker
# Spin up the OS Gateway and Vector Memory DB
make docker-up
Kubernetes
We provide official Helm charts in deploy/k8s/helm.
helm install amberclaw ./deploy/k8s/helm -f deploy/k8s/helm/values.yaml
Project Structure
AmberClaw adheres to strict domain-driven design:
src/amberclaw/agent/: ReAct planning, delegation, and learning logic.src/amberclaw/api/: FastAPI Gateway routing and lifecycle.src/amberclaw/heartbeat/: Continuous async process monitoring.src/amberclaw/memory/: 3-layer RAG and session persistence.src/amberclaw/security/: Sandboxing and egress controllers.src/amberclaw/governance/: Auto-budget limits and approval boards.