NanoClaw
Your personal Claude assistant. Single-process, minimal files, zero complexity. Secure container isolation, WhatsApp integration, and AI-native configuration via Claude Code.
Simplicity Meets Security
Single process. Minimal files. No microservices, no message queues. Just your personal Claude assistant, running securely in a container.
Container Isolation
Agents run in Linux containers with filesystem isolation. Access is restricted to mounted directories, and bash commands execute safely inside the container, never touching the host system.
WhatsApp Interface
Chat with Claude via WhatsApp groups. Each group maintains an independent context, complete with a personalized CLAUDE.md memory file.
Per-Group Memory
Every WhatsApp group has its own context and dedicated memory file. SQLite-based message storage supports per-group queuing and concurrency control.
Scheduled Tasks
Run recurring jobs using the built-in task scheduler. Automate daily reports, reminders, and background operations without external dependencies.
Claude Agent SDK
A streaming agent container powered by the Claude Agent SDK. Direct access to Claude's capabilities with built-in web search and content retrieval.
Agent Swarm
Collaborative multi-agent workflows for complex tasks. Coordinate multiple Claude instances to solve intricate problems efficiently.
NanoClaw vs. Cloud AI: Privacy and Cost Comparison
Why creators are migrating to local inference
Data Privacy
Cloud AI requires handing over conversations, files, and context to third parties. NanoClaw keeps your workflows within an environment you control.
Cost
Single-process and minimal files reduce maintenance costs, eliminating the need for complex microservices, message queues, or multi-tier backend services.
Device Support
Suitable for self-hosted environments including Linux, macOS, Windows WSL2, Docker, and Raspberry Pi.
Built for Real-World Use
From personal productivity to professional workflows, NanoClaw adapts to your needs.
Code Assistant
Local code completion, debugging, and documentation generation. Keep your proprietary code private while receiving AI assistance.
Content Creation
Draft blog posts, emails, and documents with AI assistance. All drafts and ideas remain on your device.
Data Analysis
Analyze sensitive business data locally. Data never leaves your infrastructure, ensuring complete confidentiality.
Personal Assistant
Manage calendars, emails, and tasks using voice commands and automation. Includes 100+ integrations.
Research & Learning
Summarize papers, answer questions, and organize knowledge bases. Perfect for students and researchers.
API Integration
Build custom applications using NanoClaw's REST API. Integrate AI capabilities into your existing workflows.
AI-Native Configuration
Claude Code handles everything: dependencies, authentication, containers, and service configuration.
System Requirements
Node.js, Docker or a compatible container environment, Anthropic API Key, and a device capable of connecting to WhatsApp.
What You Get
Single-process service, container isolation, per-group memory, scheduled tasks, and WhatsApp group access.
Three Simple Steps
Clone the repository, prepare your configuration, and let Claude Code automatically handle dependencies, authentication, containers, and service setup.
FAQ
What is NanoClaw?
NanoClaw is a lightweight personal Claude assistant focused on single-process, self-hosted, container-isolated operation with WhatsApp group interaction.
How much does NanoClaw cost?
The project itself is open-source; you are responsible for your own Anthropic API, server, or local hardware running costs.
What hardware is required?
A standard VPS, personal computer, Docker host, or Raspberry Pi can serve as a starting point, depending on your workload.
Can it be used without the internet?
Core configuration can be self-hosted, but the Claude API and WhatsApp integration still require an internet connection.
Where can I find the GitHub repository?
The repository is located at github.com/gavrielc/nanoclaw.
How do I update the model?
Simply update your Anthropic model configuration or Claude Code settings; there is no need to rebuild complex service clusters.