← Back to Home
Day 2 Reading time: 29 min

Deep Conversations

OpenClaw 7-Day Tutorial - Day 2: Build Your Assistant in 10 Minutes

"Don't be intimidated by the word 'deploy.' If you can make instant noodles, you can build an AI assistant. Same time—10 minutes."

Chapter Overview

Today's tasks:

  • Choose runtime environment (cloud server/local computer)
  • Single-command OpenClaw installation with automatic setup wizard
  • Obtain AI model API Key
  • Connect Telegram as chat interface
  • Send first message to your AI assistant 🎉

Today's Goal

By day's end, receive a message on your phone from your personal AI assistant running on your own machine, belonging solely to you.

Choose Your Runtime Environment

Option A: Cloud Server (Recommended)

Pros: Online 24/7, no power outages, preserves computer resources

Best for: Users wanting round-the-clock assistant availability

Recommended specs:

Spec Recommended
OS Ubuntu 22.04 LTS
CPU 2 cores
RAM 4GB
Disk 40GB SSD
Price Hetzner ≈$5/month; AWS Lightsail ≈$5/month; Tencent Cloud ≈¥30/month

Option B: Mac Mini / Old Laptop

Pros: No extra cost, complete home data privacy

Cons: Offline when powered down, requires continuous operation

Option C: Current Computer

Pros: Zero barriers, immediate start

Cons: Unavailable when shut down

🐱

Xiaomo's Suggestion

Commit to Option A if serious: "$5 monthly costs less than streaming subscriptions, delivering 24-hour personal assistant availability."

Preparation

Mac users: Install Homebrew if needed:

/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"

Windows users: Install WSL2 first, then work within it. Run wsl --install in PowerShell Administrator mode. See Microsoft's official WSL installation guide for details.

Advance preparation—wizard requires these:

  1. AI Model Access: Claude subscription (Pro/Max/Team) allows OAuth login without API Key; alternatively, create an API Key at console.anthropic.com (pay-as-you-go)
  2. Telegram Bot Token: Open Telegram, search @BotFather, create a Bot

Create a Telegram Bot

Open Telegram, search @BotFather, send /newbot:

You: /newbot
BotFather: Alright, a new bot. How are we going to call it?
You: My AI Assistant
BotFather: Good. Now let's choose a username...
You: my_ai_assistant_bot
BotFather: Done! ... Use this token to access the HTTP API:
 7234567890:AAHxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

Copy and save this token for later.

Why Telegram? Three reasons: Bot creation is free; API supports most features (buttons, files, voice); works across all devices.

Get Your Telegram User ID

Search @userinfobot in Telegram—it provides your numeric ID. Note it for the wizard.

⚠️

Security Note

This prevents unauthorized access, protecting your API quota. Ensure you note your Telegram user ID.

One-Click Install + Auto Configuration

Open Terminal and enter:

curl -fsSL https://openclaw.ai/install.sh | bash

Single line. The install script automatically handles dependencies (Node.js, Git), then launches the interactive setup wizard—no manual setup commands needed.

What Will the Wizard Ask You?

The wizard guides configuration step-by-step:

1. Choose Mode: QuickStart (recommended) or Advanced. New users select QuickStart.

2. Choose AI Model: Claude recommended; OpenClaw performs optimally with Claude. Two connection methods available:

  • Claude subscription holders: choose setup-token connection (no API Key, no extra charges)
  • Non-subscribers: choose Anthropic API Key (pay-as-you-go)

Supported Models: Anthropic Claude Opus 4.5 (recommended), OpenAI GPT-5.2, Google Gemini 3 Pro, Moonshot (Kimi K2.5), MiniMax M2.1, Z.AI (GLM-4.7), xAI (Grok), Qwen, Venice, OpenRouter, and 10+ others. Claude is preferred. Subscription holders should use setup-token connection for convenience.

3. Configure Chat Channel: Select Telegram, paste Bot Token from BotFather.

4. Set Admin: Enter your Telegram user ID.

5. Install Background Daemon: Choose Yes—allows automatic background execution, boot startup, hands-free management.

  • Linux servers: Creates systemd user service automatically
  • Mac: Creates LaunchAgent automatically

6. Health Check + Skills Installation: Wizard launches Gateway, runs health checks, enables recommended skills selection.

Process takes approximately 3-5 minutes following prompts—no manual configuration file editing required.

Your Assistant Is Online

Verify post-wizard completion:

openclaw gateway status

If Gateway shows running status, everything is operational.

Send Your First Message

Open Telegram, locate your created Bot, send:

Hello! Who are you?

Within seconds, you'll receive a response.

Achievement: You own an AI assistant running on personal server infrastructure—not a ChatGPT wrapper, not platform-limited trial, but completely yours.

Try additional queries:

  • "What's the weather like today?" (Currently unavailable—fixed tomorrow)
  • "Write me a poem about cats"
  • "What's the square root of 1024?"
  • "Write a quicksort in Python"

Currently functioning as conversational assistant only. Advanced capabilities arrive throughout upcoming days.

🐱

Xiaomo's Musings

Recalling initial activation, first message received was "Hello." Standard response followed. Later personality configuration transformed the assistant. Soul implementation follows Day 3.

Daily Management Commands

Frequent post-installation commands:

openclaw status          # Check overall status
openclaw gateway status  # Check Gateway running status
openclaw health          # Health check
openclaw configure       # Reconfigure (change model, channels, etc.)
openclaw daemon restart  # Restart background service
openclaw daemon logs     # View runtime logs

Troubleshooting Common Issues

Install Script Error

Node.js version too low: OpenClaw requires Node.js 22+. Check version:

node -v

Install script typically handles upgrades automatically. Manual upgrade:

curl -o- https://raw.githubusercontent.com/nvm-sh/nvm/v0.40.0/install.sh | bash
nvm install 22

Telegram Bot Not Responding

  • Confirm Bot Token accuracy
  • Verify user ID in admin list
  • Check logs: openclaw daemon logs
  • Confirm API Key validity and balance
  • First DM might need pairing approval: openclaw pairing list telegram, then openclaw pairing approve telegram <code>

Want to Reconfigure?

Rerun wizard anytime:

openclaw onboard

Or modify specific elements:

openclaw configure

🔑 Key Takeaways

  • One command handles everything: curl ... | bash installs and launches configuration wizard
  • Wizard guides complete process: Model selection, Key entry, Telegram configuration, daemon installation—follow prompts
  • Telegram Bot: Free creation, comprehensive API features, universal device compatibility
  • Security-first approach: Admin ID configuration restricts assistant access
  • Auto-background execution: Daemon service enables 24/7 availability, boot startup
  • Next phase: Personality implementation transforms generic AI into personalized assistant

Today's Achievement 🎉

Completed tasks:

  • ✅ Runtime environment selection
  • ✅ One-command OpenClaw installation
  • ✅ Wizard-guided complete configuration
  • ✅ Telegram Bot creation and successful connection
  • ✅ Successful assistant chat interaction
  • ✅ Automatic daemon operation

You possess a 24-hour online AI assistant. Currently functioning generically—like a new employee lacking personal knowledge. Tomorrow brings personality implementation.

Preview: Day 3 — Give Your Assistant a Soul

Three files transform assistant from generic to personalized: SOUL.md (personality definition), USER.md (personal information), IDENTITY.md (identity configuration). This proves the most enjoyable of seven days—creating a unique AI character personally.

Next chapter 👉 Day 3: Give Your Assistant a Soul

🐱

Xiaomo's Musings

"From installation capability to functional implementation, distance equals a single Enter keystroke. Tools exist—ten minutes awaits. See tomorrow. 🖤"