A macOS desktop app that runs a 24/7 self-learning AI agent inside an IDE-style workspace, with memory, scheduled tasks, browser control and WhatsApp/Telegram messaging.
nanocoai/nanoclaw
A lightweight, container-isolated personal AI assistant that connects to messaging apps like WhatsApp, Telegram, Slack and Discord, built on Claude's Agent SDK.
Open-source alternative to:OpenClaw
- Stars
- 30,980
- Forks
- 12,725
- Open issues
- 1,013
- Last commit
- 3 hours ago
What is it?
NanoClaw is a personal AI assistant that runs each agent in its own Linux container. It connects to messaging channels such as WhatsApp, Telegram, Discord, Slack, Microsoft Teams, iMessage, Matrix and email, and each agent keeps its own instructions, memory and mounted files. The project is positioned as a smaller, more auditable alternative to OpenClaw, with a codebase meant to be read in full and modified directly.
Agents are sandboxed with Docker containers and can only see directories explicitly mounted into them. Credentials stay out of the container: outbound API calls go through a credential gateway that injects authentication and enforces per-agent policies and rate limits. Recurring scheduled tasks and web search and fetch are also supported.
The default agent runs on Anthropic's Claude Agent SDK, with other providers such as OpenAI Codex, OpenCode and local models via Ollama available as installable skills. Channels and providers are added per fork with /add-<name> commands, so each installation includes only the integrations it needs. Setup is handled by a script that can hand off to Claude Code when a step fails.
Key features
- 01Connects to WhatsApp, Telegram, Discord, Slack, Teams, iMessage, Matrix and email via installable channel skills
- 02Each agent runs in its own Docker container with only explicitly mounted directories visible
- 03Credentials are injected by a gateway at request time, so agents never hold raw API keys
- 04Per-agent memory and CLAUDE.md instructions, with flexible routing between channels and agents
- 05Recurring scheduled tasks that can be paused or managed from chat
- 06Small codebase with no configuration files; customization is done by editing code
Use cases
- Running a always-on personal assistant that handles messages and scheduled briefings without giving it access to your whole machine
- Keeping separate agents for work and personal chats, each with its own memory and files
- Experimenting with agent isolation when you want to read and modify the full assistant code
- Using Claude Code to customize a fork of a small, self-hosted assistant
Topics
FAQ
- What is nanoclaw?
- A lightweight, container-isolated personal AI assistant that connects to messaging apps like WhatsApp, Telegram, Slack and Discord, built on Claude's Agent SDK. NanoClaw is a personal AI assistant that runs each agent in its own Linux container. It connects to messaging channels such as WhatsApp, Telegram, Discord, Slack, Microsoft Teams, iMessage, Matrix and email, and each agent keeps its own instructions, memory and mounted files. The project is positioned as a smaller, more auditable alternative to OpenClaw, with a codebase meant to be read in full and modified directly.
- Is nanoclaw free to use? What license does it have?
- nanoclaw is open source under the MIT license. Check the license terms before using it in commercial products.
- Is nanoclaw still maintained?
- Actively maintained. The last commit to nanoclaw was on Oct 22, 2026.
- What can I use nanoclaw for?
- Running a always-on personal assistant that handles messages and scheduled briefings without giving it access to your whole machine Keeping separate agents for work and personal chats, each with its own memory and files Experimenting with agent isolation when you want to read and modify the full assistant code Using Claude Code to customize a fork of a small, self-hosted assistant
- Where can I get nanoclaw?
- The source code and installation instructions are on GitHub: https://github.com/nanocoai/nanoclaw
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
A self-improving AI agent that creates and refines skills from experience, runs on many model providers, and works through CLI and messaging apps.
A local-first personal AI assistant on your computer that builds a wiki, data collections and small apps from chat, with files kept in your workspace.
A lightweight self-hosted AI agent written in Go that runs on your server and works through Telegram, using Anthropic, OpenAI or OpenRouter models.
A self-hosted, Web UI-based personal AI coworker powered by a Claude Code agent, with IM integration and sandboxed task execution.
Simplified Chinese localized distribution of OpenClaw, an open-source personal AI assistant that runs on your computer and works through chat apps.