English

← Back to list

Practice guide! OpenClaw how to build an AI smart team.

itadol5j

Still worried about the cumbersome and repetitive work? Want to have your own AA team? OpenClaw helps you achieve it! Today's lesson, hand-held hands teach you to build an AI team with OpenClaw.

What's OpenClaw?

OpenClawIt's an open-source AIAgent framework, which you can interpret as an "AI Auxiliary Platform".

It helps you:

  • :: Automatic mail processing
  • Automatically search for information
  • Auto-operated computers
  • Automatically respond to messages
  • Automatically manage the calendar
OpenClaw

Preparation: installation of OpenClaw

Before you start, you need to prepare:

  • One computer (Mac/Windows/Linux available)
  • Node.js environment (recommendation V18 above)
  • OpenAI API Key (or other large model API)

Installation steps:

1. Open terminal (Mac/Linux) or command hint (Windows)

2. Cloning projects:

git clone https://github.com/openclaw/openclaw.git
cd openclaw
npm install

3. Profile:

Copy profile template:

cp config.example.yaml config.yaml

4. Edit profile to fill in your API key:

api:
  openai:
    api_key: your-api-key-here

Start-up services:

npm start

After starting successfully, access http://localhost:8080 can open the management interface.

Step 1: Create your first AI smart body

After entering the management interface, click the Create Smart button.

You need settings:

  • Name: Name your AI.
  • Description: Tell AI what it does.
  • Skills: Select the capabilities it needs (mail, calendar, search, etc.)
  • Character: Set its response style

Create examples:

名称：邮件助手
描述：帮我自动处理邮件，分类重要邮件，回复常规邮件
技能：读取邮件、发送邮件、标记重要
性格：专业、简洁

Step 2: Skills and workflow configuration

OpenClaw’s strength is that it can be equipped with “skills”.

Common skills include:

Skills Functions Apply scene
Mail skills Read/send/classify mail Automatically process mail
Calendar skills Create/view/modify the schedule Time management
Search skills Network search, information capture Data collection
Browser skills Operation Browser, Filling Table Automation of web operations
Documentation skills Read and write documents, organize information Document Management

Configure workflow:

Workstream = automated task combining multiple skills.

For example, the daily briefing stream:

  • Search skills
  • Documentation skills Trail
  • 3. Mail skills

Step 3: Dialogue with AI smart bodies

When the configuration is complete, you can talk to the AI smart body!

Mode I: Web Interface

Talk directly on the management interface, as simple as using Twitter.

Mode 2: Integration into communication tools

OpenClaw supports integration to:

  • Telegram
  • Discord.
  • WhatsApp
  • Flying Book
  • Email

Configure examples (as in Telegram):

在 config.yaml 中配置：
telegram:
  enabled: true
  bot_token: your-bot-token
  chat_id: your-chat-id

Progress: Building an AI smart team

One AI isn't enough? You can create several professional AI teams!

Example team configuration:

  • Researcher: search for information, organize information
  • The writer.: writing articles, generating content
  • • Analyst: responsible for data analysis, generation of reports
  • Client service: responding to users and solving problems

Team working model:

你：帮我写一篇关于AI的科普文章

研究员：搜索AI相关资料...
写手：根据资料撰写文章...
分析师：检查文章质量...

最终输出：完整的科普文章

• Field cases: automatic weekly reporting generator

Let me show you a practical case: Autoweekly.

Step 1: Create an intelligent body

名称：周报助手
描述：每周自动汇总工作内容，生成周报
技能：日历查询、文件读取、文档生成

Step 2: Setting trigger conditions

It is automatically triggered at 5 p.m. on Fridays or manually entered "to produce a weekly newspaper".

Step 3: Workstream

  • 1. Query this week ' s calendar
  • 2. Read the working documents
  • 3. AI generation → organized into weekly reporting formats
  • Send mail to your mailbox.

Effects: A full weekly report is automatically received every Friday.

⚠️ Common answer

Q1: Do you need a programming base?

A: Basic configuration is not required, computer is available. Advanced customization requires a little programming knowledge.

Q2: API going to pay?

A: OpenClaw is free of charge per se, but the call to the large model API requires a fee (at the rate of call).

Q3: Is it safe? What about privacy?

A: Local processing of data to support self-deployment, fully manageable.

Q4: Do you support Chinese?

A: All in Chinese!

Summary

OpenClaw enables everyone to have their own AI smart team:

  • Zenium Install SimpleA few lines of command.
  • Zenium Flexible configuration: Take whatever you want.
  • Zenium Free open source: Don't waste money
  • Zenium Chinese friendly: national user preferences

Don't wait to build your AIS team today!

Action guide: Install → Create AI → Configure Skills → Start using