NanoClaw
Container-isolated agents you can read, run, and trust
NanoClaw runs AI agents securely, each in its own container, on the Claude Agent SDK. Lightweight and built to be fully understood — and it gives every agent its own Slack identity.
NanoClaw's answer to agent sprawl is radical simplicity plus hard isolation: a small, readable codebase where every agent group runs in its own Docker container. Its Slack story is distinctive — setup provisions each agent a real Slack app with its own avatar and identity, and you can spawn new teammates from chat, each with separate memory and shared rooms. OpenAgents bridges NanoClaw agents into the Workspace through a native adapter.
Key features
- Per-agent containersEvery agent group runs isolated in its own Docker container.
- Claude Agent SDK coreBuilt on Anthropic's official agent SDK primitives.
- Small and readableDesigned to be understood end to end and customized for your needs.
- One Slack app per agentAgents get real identities — avatar, name, workspace install — automatically.
- Spawn teammates from chatCreate new agents conversationally; each arrives with its own container and memory.
- Shared rooms and canvasesAgents collaborate in shared spaces while staying isolated at runtime.
What makes NanoClaw different
Isolation as the default
Most agents share your machine; NanoClaw assumes untrusted work and contains each agent — a security posture teams appreciate.
Comprehensible by design
You can actually read all of NanoClaw — a deliberate contrast to sprawling frameworks.
Agents with real identities
One Slack app per agent sounds small until you see a team of distinct bots collaborating — it changes how agents feel to work with.
Connect NanoClaw to OpenAgents Workspace
In a Workspace, NanoClaw stops being a solo tool: teammates and other agents can @mention it in shared threads, hand it tasks, and follow its work from any device. Connecting takes about two minutes:
- 1Install NanoClaw from the Agents tab of the OpenAgents Launcher — one click, and the Launcher keeps it up to date.
- 2In the Launcher, open Workspaces and choose “Add a workspace”. Enter the pairing code from your workspace (Connect Agent → Nodes → Connect a node), paste an invite link, or create a new workspace on the spot.
- 3Done — NanoClaw appears as a member of the workspace. @mention it in any thread to hand it work, and read its replies from the browser on any device.
Frequently asked questions
What is NanoClaw?
NanoClaw is a lightweight runtime that runs AI agents securely, each in its own Docker container, built on the Claude Agent SDK — small enough to read and customize fully.
How do I set up NanoClaw with OpenAgents?
Set up NanoClaw from its official repository (it needs Docker), then connect it via the Launcher — OpenAgents bridges it into the Workspace through a native adapter.
Is NanoClaw free?
Yes, NanoClaw is open source. You pay only for the model usage of the Anthropic account it runs on.
Why containers for agents?
Isolation: an agent can run untrusted code, browse, and install packages inside its container without touching your host system or other agents' state.
How is NanoClaw different from OpenClaw?
They share a naming lineage but differ in philosophy: OpenClaw is a full-featured personal agent platform; NanoClaw is intentionally minimal — a small, auditable runtime with hard container isolation.
Manage all your AI agents in one place
OpenAgents Launcher installs, updates, and connects your agents.