AI/TLDR

NanoClaw

A self-hosted personal assistant whose agents each run in their own container

Assistants & ChatbotsOpen source
Language
TypeScript
License
MIT
$git clone https://github.com/nanocoai/nanoclaw.git nanoclaw-v2

Overview

NanoClaw is a self-hosted personal AI assistant that reaches you through the messaging apps you already use, and its distinguishing choice is where the sandbox lives. Rather than restricting what an agent may do with application-level permission prompts, NanoClaw runs each agent inside its own Linux container with explicit mount controls, so the isolation is enforced by the operating system. The README positions it as a lighter, more legible alternative to OpenClaw — "small enough to understand", with few dependencies.

Underneath it is built on Anthropic's official Agent SDK, and it adds the pieces a long-lived personal assistant needs: persistent conversation memory, recurring scheduled jobs, and credential handling that routes API access through OneCLI's Agent Vault so raw keys never enter a container. User data, credentials and agent state stay on your machine by default.

Channel coverage is broad — WhatsApp, Telegram, Discord, Slack, Microsoft Teams, iMessage, Matrix, Google Chat, Webex, Linear, GitHub, WeChat and email via Resend — so the same agent can be reached wherever a conversation happens to start. The runtime is TypeScript and Node with Bun, and the project is MIT-licensed. Containers require Docker on macOS, Linux or WSL2.

What it does

  • OS-level isolation: every agent runs in its own Docker container with explicit filesystem mount controls
  • Built on Anthropic's official Agent SDK
  • Messaging on WhatsApp, Telegram, Discord, Slack, Microsoft Teams, iMessage, Matrix, Google Chat, Webex, Linear, GitHub, WeChat and email
  • Persistent conversation memory plus recurring scheduled tasks
  • Credentials brokered through OneCLI's Agent Vault, keeping raw API keys out of containers
  • Deliberately small codebase with minimal dependencies, meant to be read end to end

Getting started

One script does the whole setup: it installs Node, pnpm and Docker if they are missing, registers your credentials, builds the containers and connects your first messaging channel.

Clone the repository

Fetch the project and change into it.

bashbash
git clone https://github.com/nanocoai/nanoclaw.git nanoclaw-v2
cd nanoclaw-v2

Run the installer

The setup script handles dependencies, credential registration, container builds and channel configuration.

bashbash
bash nanoclaw.sh

Connect a channel

During setup you pick your first messaging channel — WhatsApp, Telegram, Slack, Discord, Teams, iMessage or email. Additional channels can be added afterwards.

Give the agent a home directory

Because agents run in containers, decide explicitly which host directories to mount before letting an agent work on your files — that mount list is the security boundary.

Commands and code are distilled from the project's own documentation — always check the official repo for the latest.

When to use it

  • Reach for it when you want a personal assistant on WhatsApp or Telegram but are not comfortable giving an agent unsandboxed access to your machine
  • Use it when you want the assistant self-hosted so conversation history and credentials never leave your hardware
  • Good for recurring jobs — a scheduled digest, a daily triage pass — that should run unattended in an isolated container
  • Useful when you want a small enough codebase to audit before trusting it with your accounts

How NanoClaw compares

NanoClaw alongside other open-source assistants & chatbots tools AI/TLDR tracks, ranked by GitHub stars.

ToolStarsWhat it does
OpenClaw★ 390kOpenClaw is a self-hosted personal AI assistant that answers you on WhatsApp, Telegram, Slack, Discord, and many other channels, with voice and a live visual canvas.
Hermes Agent★ 246kA self-improving personal AI agent from Nous Research that builds skills from experience, remembers across sessions, and reaches you on Telegram, Discord, Slack, and more.
Odysseus★ 87.3kA self-hosted AI workspace that puts chat, agents, deep research, documents, email, notes, tasks and calendar behind one Docker Compose stack, over local or API models.
CowAgent★ 47kA self-hosted assistant that plans and executes tasks with built-in file, terminal, browser and search tools, and answers across a web console plus a dozen messaging platforms.
AstrBot★ 40.6kAn all-in-one agent chatbot platform that puts LLM conversations, tools, knowledge bases and a plugin marketplace inside messaging apps like Telegram, Slack, Discord, QQ and Feishu.
OpenHuman★ 39.8kA local-first desktop personal AI for macOS, Windows and Linux that keeps a compressed memory tree on your machine and orchestrates checkpointed research and automation workflows.
MindsHub★ 39.7kAn agent workspace for knowledge work and software development that runs swappable open-source agent harnesses against your choice of frontier or open models.
NanoClaw★ 30.8kA self-hosted personal assistant whose agents each run in their own container