Deploy Clawdbot: Step‑by‑Step Guide to Your Personal AI Agent

Tool Overview

Clawdbot is an open‑source personal AI agent that runs locally, can be controlled via chat platforms such as Telegram or WhatsApp, and automates tasks like file organization, code deployment, and calendar sync. By January 2026 it has over 40 k GitHub stars and supports Windows, macOS, Linux and cloud servers.

Prerequisites

Hardware

CPU: minimum 4‑core 64‑bit, recommended 6‑core (Intel i5 / AMD Ryzen 5 or higher).

Memory: at least 8 GB; 16 GB+ recommended for local large‑model inference.

Storage: 10 GB free, 20 GB+ recommended.

Network: stable internet, 100 Mbps+ for API access.

Software

Node.js ≥ 22.0.0 (install from the official site).

Optional NVM for version management.

Docker ≥ 20.10.0 (recommended).

Git (any stable version).

Operating system: Windows 10 2004+/11 with WSL2, macOS 12+ (Intel/Apple Silicon), or Debian/Ubuntu 20.04+ Linux.

Administrator / sudo privileges.

Required Credentials

AI model API key (Claude, GPT‑4, GLM‑4.5, etc.).

Chat platform token (Telegram Bot Token, WhatsApp/Discord/Slack token).

Download Sources

GitHub repository (clone or download zip).

One‑click install scripts: curl -fsSL https://clawd.bot/install.sh (Mac/Linux) or iwr -useb https://clawd.bot/install.ps1 (Windows PowerShell).

Plugin marketplace “Clawdhub”.

Official documentation on GitHub Docs.

Installation Methods

Method 1 – Docker Deployment (recommended for beginners)

Install Docker and verify with docker --version. Windows users enable WSL2 and install Ubuntu 22.04 LTS.

Clone the repository and copy the example environment file:

git clone https://github.com/clawdbot/clawdbot.git
cd clawdbot
cp .env.example .env
vim .env

Modify required variables such as CLAWS_BOT_NAME, CLAWS_SERVER_PORT, CLAWS_DEFAULT_MODEL, and CLAWS_DATA_DIR.

Build and run the container: docker-compose up -d --build Check status with docker ps and view logs via docker logs clawdbot.

Open http://localhost:8080 (or the configured port) in a browser to confirm the management UI.

Method 2 – One‑Click Script (Mac/Linux)

Run the official script: curl -fsSL https://clawd.bot/install.sh | bash Launch the onboarding wizard: clawdbot onboard --install-daemon Follow prompts to select a start mode, provide an API key, and choose a chat channel.

Method 3 – Global npm Installation (advanced)

Ensure Node 22+ (use NVM if needed):

nvm install 22
nvm use 22

Install Clawdbot globally: npm install -g clawdbot@latest or pnpm add -g clawdbot@latest Run the onboarding wizard: clawdbot onboard Start the gateway service:

clawdbot gateway --port 18789 --verbose

Telegram Integration

Obtain a Bot Token from @BotFather.

Add the token to the .env (Docker) or ~/.clawdbot/clawdbot.json (script) as TELEGRAM_BOT_TOKEN="your_key".

Restart the service ( docker-compose restart or clawdbot gateway restart).

Test by sending /start to the bot and issuing a command such as “list current directory files”.

Common Commands

File management: “整理下载文件夹,按类型分类 PDF、图片、文档”.

Code execution: “运行当前目录下的test.py脚本,返回结果”.

Calendar sync: “将明天 14:00 的会议添加到系统日历,同步至 Notion”.

Text processing: “将当前目录的会议录音转成文字纪要”.

System monitoring: “监控服务器 CPU 使用率,超过 80%提醒我”.

Advanced Configuration

Connecting a Chinese model (GLM‑4.5)

Register on the Zhipu AI platform and obtain an API key.

Edit ~/.clawdbot/clawdbot.json to add a provider entry for glm with the base URL https://open.bigmodel.cn/api/paas/v4 and your API key.

Restart the gateway.

Installing Skill Plugins

Download desired plugins from Clawdhub.

Place them in the clawdbot/plugins directory.

Activate with clawdbot plugin install.

Troubleshooting

Installation failures

Docker version too low – upgrade to 20.10+ and run docker-compose down -v before retrying.

Node.js version below 22 – install via NVM.

Clone failures – download the zip archive or configure a proxy.

Functional issues

Telegram bot unresponsive – verify the token, add PROXY_URL for mainland China.

File permission errors – map host directories and set chmod 777 on them.

Performance lag – disable local large models or switch to a smaller cloud model (e.g., llama3:7b).

Safety notes

Clawdbot runs with system‑level privileges; avoid installing on a primary workstation.

Prevent sleep mode and, for WSL2, run wsl --shutdown to avoid interruptions.

For cost‑effective usage, prefer domestic models for light workloads and Claude Opus for heavy usage.