OSS Tanbou
View on GitHub

NanoClaw — A lightweight personal AI assistant that isolates agents in containers and connects messaging, memory, and scheduled work

OSS health 96
About these scores

Discovery score is an unbounded weighted, log-compressed index of stars, watchers, forks, and contributors. Growth momentum is its change over the observed period; OSS health is a 0–100 score from available repository recency, Community Health, and release data.

Stars
30,792
Primary language
TypeScript
License
MIT
Repository last updated
Sep 17, 2026

Overview

NanoClaw uses a host router to receive messages and launch each agent in its own Linux container, running Claude Agent SDK or optional providers. Channels such as Slack, Telegram, Discord, and WhatsApp are installed as skills, alongside memory, scheduled tasks, MCP tools, and provider extensions. The latest release checked on September 17, 2026 is v2.3.0.

Features and best fit

Based on official documentation; not hands-on tested · Content checked:

Reduce host exposure with per-agent containers and explicit filesystem mounts

Instead of relying only on application-level allowlists, NanoClaw isolates agent groups into separate containers and workspaces. OneCLI Agent Vault injects credentials at request time so raw API keys do not need to live in the agent container.

Sources: [2]

Install channels and providers as skills, then add memory and scheduled jobs per agent group

Messaging adapters are added only when needed, and provider skills can switch agents to Codex, OpenCode, or Ollama. Agent groups can carry their own memory, MCP tools, skills, and scheduled tasks.

Sources: [2]

For individuals who want powerful personal agents while keeping the codebase and security boundary understandable

It fits users who are comfortable maintaining a fork and want messaging automation with explicit filesystem isolation and credential policy per agent.

Sources: [2]

v2.3.0 contains several breaking migrations and now requires Node.js 22+

The release changes the session runtime driver, central DB seam, agent mailbox access, and Slack agent model. Existing forks should run the documented detection/migration paths. Container isolation also does not make mounted directories safe by itself, so mounts and tool permissions still need review.

Sources: [3][2]

Official sources

  1. [1]nanocoai/nanoclaw — GitHub repository metadata(2026-09-17)
  2. [2]NanoClaw — README(2026-09-17)
  3. [3]NanoClaw v2.3.0 release(2026-09-17)
Supplemental curator note

Its defining choice is per-agent container isolation rather than relying only on application-level permission checks.

Try it in 3 steps

  1. 1

    Get the source

    git clone --depth 1 https://github.com/nanocoai/nanoclaw.git
  2. 2

    Enter the repository

    cd nanoclaw
  3. 3

    Check the official steps

    Continue with the commands in the README Installation, Quick Start, or Getting Started section.

    find . -maxdepth 1 -iname 'README*' -exec sed -n '1,220p' {} \; -quit
Check the official README

Growth

Growth trends · Last 30 days

30,792 Stars

Trend data is still being collected.

Development activity

Last 90 days · weekly

Commits (last 30 days)
330
Open PRs
700

Development activity is still being collected.

Built with

Categories and tags

Categories

Tags

GitHub data

GitHub dataView detailed GitHub data
View on GitHub

GitHub Topics

  • ai-agents
  • ai-assistant
  • claude-code
  • claude-skills
  • openclaw
Stars
30,792
Forks
12,831
Watchers
134
Open issues
359
Primary language
TypeScript
License
MIT
Repository last updated
Sep 17, 2026
Write a related article

Share a guide or use case for this OSS in Markdown. Articles are published after administrator approval.

Report incorrect information

Tell us if any listing information is incorrect or outdated.