🇹🇼 中文 🇺🇸 English 🇯🇵 日本語 🇪🇸 Español
🦞 808AI 五金行

The Complete OpenClaw Setup Guide

From beginner to advanced, cloud to local — everything you need to raise your OpenClaw

01

What is OpenClaw?

OpenClaw is an open-source AI Agent framework that lets you run a team of AI workers on your own machine. Command them remotely via Telegram, Discord, LINE, and more. They can:

In short: you give the orders, OpenClaw gets things done. 24/7, no breaks.
02

OpenClaw vs Other Solutions

There's more than one AI Agent solution out there. Which one is right for you?

⚔️

Claude Code
Solo Operator

Use Claude Code CLI directly — ask and receive. Great for short tasks, but no persistent memory, no scheduling, and it's gone when you close it.

🏢

Manus / Coze
Managed Platforms

The platform handles everything — convenient but limited. Low customizability, your data lives on someone else's cloud, and features are capped by the platform.

03

Model Selection Guide

Different models have different strengths. Pick the right one and work smarter.

Claude Sonnet Claude Opus GPT-4o Gemini 2.5 Pro
RoleDaily workhorseHighest qualityGeneral purposeLong context
SpeedFastSlowerFastFast
CostMediumHighMediumMedium
Best forEveryday tasksComplex reasoning / writingMultimodalLarge datasets
💡
808AI Recommendation
  • Use Sonnet for daily tasks, Opus for important or complex ones
  • OpenClaw supports model failover — if one goes down, it automatically switches to the next
04

How Much Does It Cost?

Setting up OpenClaw has two main cost areas: hardware and API.

Hardware Cost (Local)

DevicePriceSuitable For
Mac mini M4 16GB~NT$20,000Entry-level, run 2–3 Agents
Mac mini M4 Pro 24GB~NT$35,000Advanced, multi-Agent + browser automation
VPS (Zeabur / Railway)~NT$0–500/moDipping your toes in

API Cost (Model Fees)

PlanMonthlyNotes
Claude Pro$20/moGood for light personal use
Claude Max 5x$100/moFor power users
API Pay-as-you-goUsage-basedMost flexible, ideal for multi-Agent setups
💰
Money-Saving Tips

Multi-Agent + model tiering can save 50%+ on token usage. Use cheap models for simple tasks, expensive ones for the important stuff.

05

Cloud vs Local

Step one: decide where OpenClaw runs.

☁️

Cloud Deployment

  • One-click deploy on Zeabur / Railway
  • No need to keep your own machine running 24/7
  • Great for testing the waters or light usage
  • Limitation: can't access local files or browser
  • Cost: monthly subscription
🧭
808AI Recommendation

Start with cloud to test the waters. Once you're committed, grab a Mac mini.

06

Six Steps to Get Started with OpenClaw

From zero to one — just follow along.

1

Prepare Your Environment

Mac / Linux / VPS all work. On Mac, press Cmd+Space and type "Terminal" to open it.

2

Install Node.js

Paste this in your terminal: curl -fsSL https://fnm.vercel.app/install | bash && fnm install --lts

3

Install OpenClaw

npm install -g openclaw

4

Initial Setup

openclaw init — follow the prompts to enter your API Key and messaging app settings.

5

Launch OpenClaw

openclaw start — when you see 🦞, you're good to go!

6

Start Chatting

Open Telegram / Discord and say hello to your AI agent!

07

Core Concepts

Understand these concepts and you'll understand OpenClaw's architecture.

🤖

Agent

Your AI worker. Each Agent has its own name, personality, tools, and memory. You can have one for development, one for marketing, one for research.

🏷️ Think of it as: specialized employees
🛠️

Skill

An instruction manual for your Agent. Install a Skill, gain an ability. The community has hundreds of ready-made Skills to install.

🏷️ Think of it as: employee training manuals
🧠

Memory

Your Agent's long-term memory. It remembers your preferences, lessons learned, and key decisions. The more you use it, the better it knows you.

🏷️ Think of it as: a work journal
💓

Heartbeat

A scheduled self-check mechanism that runs automatically. Like an alarm clock that periodically triggers your Agent to inspect, report, and maintain.

🏷️ Think of it as: daily standup meetings
🔌

MCP (Model Context Protocol)

The USB-C of the AI world. Lets your Agent connect to external tools like Notion, GitHub, Figma, and more.

🏷️ Think of it as: a universal adapter

Cron (Scheduled Tasks)

Set a time, and your Agent runs automatically. Daily reports at 9 AM, weekly backups on Friday — all automated.

🏷️ Think of it as: a shift schedule
08

Day-to-Day OpenClaw Operations

Common commands and config files — bookmark this for quick reference.

/statusCheck system status
/modelSwitch models
openclaw startStart OpenClaw
openclaw stopStop OpenClaw
openclaw restartRestart OpenClaw
MEMORY.mdAgent long-term memory file
HEARTBEAT.mdHeartbeat check configuration
AGENTS.mdAgent role definitions
09

Advanced Techniques

Once OpenClaw is up and running, unlock more powerful capabilities.

👥

Multi-Agent Collaboration

Run multiple agents, each with their own role: one for development, one for marketing, one for research. Independent memory, independent models, no interference.

🌐

Browser Automation

Let your Agent control the browser: search the web, fill forms, take screenshots, scrape data — like having an assistant sitting right next to you.

💻

Coding Agent

Connect Claude Code / Codex and let your Agent write code, open PRs, and run tests automatically. All you do is merge.

📱

Messaging App Integration

Telegram / Discord / LINE / Slack — command your AI team from your phone, anytime, anywhere.

🧬

Self-Evolution

Through Self-Reflection, Memory, and Insight mechanisms, your Agent learns from mistakes and gets stronger over time.

10

Five Principles for Running OpenClaw

Wisdom from experienced OpenClaw users — skip the common pitfalls.

01

Get It Running First, Optimize Later

Don't chase the perfect setup from day one. Install it, start using it, and fine-tune as you go.

02

One Agent, One Job

Don't make a single Agent do everything. Clear division of labor means higher efficiency.

03

Feed Your Memory Well

Your Agent's quality depends on the quality of memory you give it. Invest time in writing solid MEMORY.md and AGENTS.md files.

04

Leverage Community Skills

Don't build everything from scratch. ClawHub has hundreds of ready-made Skills. Stand on the shoulders of giants.

05

Watch Your Costs

API fees add up. Use model tiering (cheap models for simple tasks, expensive ones for the important stuff) to keep costs in check.

11

Community Curated Resources

High-quality tutorials and shares from the community, continuously updated.

12

808AI 五金行 Recommended Builds

Pick what fits your needs — we've put together the packages for you.

🔧

Starter Kit

  • Zeabur Free Plan
  • Claude Sonnet API
  • Telegram Bot
  • 1 Agent
⚙️

Pro Kit

  • Mac mini M4 Pro
  • Multi-model failover (Claude + GPT + Gemini)
  • 10+ Agent team
  • Full Skill ecosystem
  • Cron scheduled automation
  • All messaging app integrations
🛡️ 🛡️ OpenClaw Security Checklist — 10 minutes to ensure your OpenClaw is secure
💬 Join the 808AI 五金行 LINE community and connect with fellow AI builders Join Community