What is NanoClaw?
NanoClaw is a remarkably lightweight AI assistant that securely runs personal agents inside their own dedicated containers. Instead of being a massive, bloated framework, it’s a minimalist tool designed to be completely understood and infinitely customized to your exact needs.
It is AI-native by design: you don’t endlessly configure it; you fork it and simply ask the AI to modify the codebase to match your workflow.
Why It’s the Perfect Alternative to OpenClaw
OpenClaw is an undeniably impressive project with a great vision, but giving complex software full access to your digital life can keep you up at night. OpenClaw relies on application-level security, like allowlists and pairing codes, running everything in a single Node process with shared memory.
It’s also huge, burdened with nearly half a million lines of code, dozens of config files, and 70+ dependencies.
NanoClaw gives you that same core magic, but in a footprint so small (around 500 lines) that you can actually comprehend it. Instead of just hiding behind permission checks, NanoClaw agents are physically isolated in their own Linux or Apple containers with strict filesystem boundaries.
It’s true OS-level isolation versus an honor system. If you want to understand everything going on under the hood, you can just ask Claude Code to read the handful of source files and walk you through it.
The Feature Lineup
- Container Sandboxing: Agents run in isolated containers (Docker, WSL2, or macOS Apple Containers) and can only see exactly what you explicitly mount for them.
- Multi-Channel Messaging: Chat with your agent on WhatsApp, Telegram, Discord, Slack, iMessage, and more. NanoClaw uses “skills over features”—you just run a command like
/add-telegramand it injects only the code you need. - Per-Agent Workspaces: Each agent group gets its own isolated memory, its own instruction set, and its own container boundary. Nothing crosses over unless you specifically wire it up.
- Scheduled Tasks: Set up recurring jobs executed by the agent, featuring smart script gates so it doesn’t wake up unless there is actual work to do.
- Credential Security (Agent Vault): Agents never hold your raw API keys. All outbound requests route through a vault that injects credentials at the time of the request while enforcing strict rate limits.
- AI-Native Customization: There are no messy UI dashboards or sprawling configuration files. If you want to debug an issue or add a feature, you just describe the problem in chat and Claude Code handles the code changes for you.
What It Supports
- Multi-channel messaging — WhatsApp, Telegram, Discord, Slack, Microsoft Teams, iMessage, Matrix, Google Chat, Webex, Linear, GitHub, WeChat, and email via Resend. Installed on demand with
/add-<channel>skills. Run one or many at the same time. - Flexible isolation — connect each channel to its own agent for full privacy, share one agent across many channels for unified memory with separate conversations, or fold multiple channels into a single shared session so one conversation spans many surfaces. Pick per channel via
/manage-channels. See docs/isolation-model.md. - Per-agent workspace — each agent group has its own
CLAUDE.md, its own memory, its own container, and only the mounts you allow. Nothing crosses the boundary unless you wire it to. - Scheduled tasks: recurring jobs executed by the agent, with optional script gates that avoid waking it when there is no work
- Web access — search and fetch content from the web
- Container isolation — agents are sandboxed in Docker containers (macOS/Linux/WSL2)
- Credential security — agents never hold raw API keys. Outbound requests route through OneCLI’s Agent Vault, which injects credentials at request time and enforces per-agent policies and rate limits.
- Agent templates: stamp a ready-to-run agent (instructions + MCP tools + skills, no secrets) from a reusable bundle via
ncl groups create --template <ref>. Templates load from the localtemplates/folder; populate it by hand or by copying from the public library. See docs/templates.md.
Why You May Want to Use NanoClaw
You should use NanoClaw if you value understanding your tools and prefer code over configuration. It’s built for the individual user who wants a bespoke personal assistant, not a monolithic piece of bloatware.
If you want an agent that is safe enough to execute local bash commands because it’s safely trapped inside a sandbox, this is for you. Ultimately, you use NanoClaw because you want an AI companion that fits your life perfectly, small enough to read in an afternoon, secure enough to trust with your data, and capable of coding its own upgrades whenever you ask it to.
License
MIT License.
