How to Set Up Clawd Bot on Your Local Machine (Mac & Ubuntu) and Connect It to WhatsApp
HOW TO GUIDES Jan. 29, 2026, 12:20 a.m.

How to Set Up Clawd Bot on Your Local Machine (Mac & Ubuntu) and Connect It to WhatsApp

Running your own AI assistant locally gives you something most cloud tools can’t: full control. With Clawd Bot, you can run a personal AI agent on your own machine, give it access to your system, and talk to it directly from chat apps like WhatsApp.This guide walks you through:

  • Installing Clawd Bot locally (Mac & Ubuntu)

  • Giving it system-level access

  • Connecting it to WhatsApp

  • Using it to control tasks and automation

Prerequisites

Before starting, make sure you have:

  • A Mac (Apple Silicon or Intel) or Ubuntu 20.04+

  • Basic terminal knowledge

  • Docker installed

  • Node.js (LTS recommended)

  • A WhatsApp account (personal, not Business API)

  • An AI model API key (OpenAI, Anthropic, or local model like Ollama)

Step 1: Install Required Dependencies

On macOS

brew install docker node git

Start Docker Desktop and make sure it’s running.

On Ubuntu

sudo apt update
sudo apt install -y docker.io docker-compose nodejs npm git
sudo usermod -aG docker $USER

Log out and log back in to apply Docker permissions.

Step 2: Clone the Clawd Bot Repository

git clone https://github.com/clawd-bot/clawd
cd clawd

This repository contains the core agent, integrations, and messaging connectors.

Step 3: Configure Environment Variables

Create a .env file:

cp .env.example .env

Edit it:

nano .env

Set:

  • AI provider (OpenAI / Anthropic / local)

  • API key

  • Memory storage path

  • Messaging adapters you want enabled

Example:

AI_PROVIDER=openai
OPENAI_API_KEY=your_api_key_here
MEMORY_BACKEND=local
ENABLE_WHATSAPP=true

Step 4: Give Clawd Bot Local System Access

This is what makes Clawd Bot powerful.You can allow it to:

  • Read/write files

  • Run shell commands

  • Schedule tasks

  • Trigger scripts

Enable permissions in the config:

ALLOW_SHELL=true
ALLOW_FILE_SYSTEM=true
ALLOW_SCHEDULER=true

⚠️ Important: Only do this on a trusted machine. This turns Clawd Bot into a real system-level assistant.

Step 5: Start Clawd Bot

Using Docker (recommended):

docker compose up -d

Or directly:

npm install
npm run start

At this point, Clawd Bot is running locally and waiting for messages.

Step 6: Connect Clawd Bot to WhatsApp

Clawd Bot connects to WhatsApp using a WhatsApp Web session (similar to how tools like whatsapp-web.js work).

Steps:

  1. Enable WhatsApp connector:

WHATSAPP_MODE=web

  1. Start the WhatsApp service:

npm run whatsapp

  1. A QR code will appear in the terminal.

  2. Open WhatsApp on your phone:

    • Settings → Linked Devices

    • Scan the QR code

Once connected, Clawd Bot appears as a chat contact you can message.

Step 7: Talking to Clawd Bot on WhatsApp

Now comes the fun part.From WhatsApp, you can send messages like:

  • “Summarize today’s calendar”

  • “Create a reminder at 9 PM”

  • “Scan my Downloads folder for PDFs”

  • “Run my backup script”

  • “Draft an email reply”

Clawd Bot processes the request locally and responds back on WhatsApp.

Step 8: Automating Tasks

You can define workflows like:

  • Auto-respond to certain messages

  • Daily summaries

  • File monitoring

  • Scheduled scripts

Example:

@clawd every day at 10am send me a summary of pending tasks

This turns WhatsApp into a command center for your system.

Security & Best Practices

  • Use a dedicated WhatsApp account if possible

  • Run Clawd Bot inside Docker for isolation

  • Avoid exposing ports publicly

  • Regularly review enabled permissions

  • Back up memory and config files

Common Use Cases

  • Personal “Jarvis” assistant

  • Developer automation helper

  • Inbox and reminder management

  • System monitoring via chat

  • AI-powered command execution

Final Thoughts

Setting up Clawd Bot locally gives you something rare:
a private, always-on AI assistant that lives inside your chats and runs on your machine.Instead of opening dashboards or apps, you simply message your assistant and get things done — very natural, very powerful, and very future-ready.


Share this article