OpenClaw is a personal AI assistant that runs on your own computer and talks to you through Telegram, Discord, or any chat app.
Pick your setup path. Both end up in the same place.
For people on the VPS path. Already SSH'd in? Skip this slide โ
SSH is a way to securely connect your Terminal to a remote computer (your VPS) as if you were sitting in front of it.
YOUR_SERVER_IP with your actual IP address:
root@your-server:~# you're in. ๐
Your server IP was shown when you created your DigitalOcean droplet. It looks like: 167.99.123.45
You can also find it in your DigitalOcean dashboard under Droplets.
OpenClaw requires Node.js. Check if you have it first:
If you see a version number (like v22.x.x or v24.x.x), skip to Step 2 โ
If you get an error, install Node.js:
โ nodejs.org/en/downloadDownload the LTS version. Run the installer. Come back here when done.
Open Terminal and run this command:
This takes about 1 minute. You'll see it download and install.
When it's done, you'll see a message that says OpenClaw is ready.
On a VPS via SSH, use the same command.
Having trouble? โ docs.openclaw.ai/start/getting-started
Now run this command to configure your agent:
The wizard will ask you:
Create an account, go to API Keys, and create a new key. Copy it.
Run this to confirm your Gateway is up:
You should see: Gateway running on port 18789
Open your dashboard:
This opens a browser window. If you see a chat interface, you're done with setup.
No response? โ docs.openclaw.ai/help
Kalibr gives OpenClaw self-healing pipelines and intelligent model routing. As your agent runs, it learns what works and automatically routes each task to the best model, getting more accurate and less expensive over time.
Free to start ยท No credit card required
Replace YOUR_EMAIL with your email address.
Your agent handles everything: account creation, plugin install, API key setup, routing config, and workspace file creation. Takes about 3 minutes. Zero manual steps.
โ Kalibr + OpenClaw setup complete.
๐ Dashboard: dashboard.kalibr.systems
๐ Add provider keys anytime: OpenAI, Anthropic, DeepSeek
Message your agent from your phone. Takes about 2 minutes.
Step 1. Open Telegram on your phone. Search for @BotFather (look for the blue checkmark โ it's official).
Step 2. Send this message to BotFather:
Step 3. Follow the prompts. When it asks for a name, call it anything (like 'My Agent'). BotFather will give you a bot token that looks like:
Step 4. When the OpenClaw setup wizard asks "Which channel?" โ choose Telegram. Paste your bot token when it asks.
Step 5. Find your new bot in Telegram (search by the name you gave it). Send it a message. Your agent will reply.
Done with setup already? Run the wizard again:
Full Telegram docs: docs.openclaw.ai/channels/telegram
Once your agent is running, these make it more capable.
Add Nebius GPU compute as a model provider in OpenClaw. Faster, cheaper inference for heavy workloads.
Plugin: clawhub.ai/plugins/tokenfactory | Guide: github.com/opencolin/openclaw-nebius
Gives your agent the ability to search the web. Ask it to research anything in real time.
Get a free API key at tavily.com
ClawHub has a growing library of skills and plugins for OpenClaw. Browse what's available and add capabilities to your agent.
Gives your agent persistent memory across conversations.
None of these are required. Your agent works great without them.
Save this. When you get home and something isn't working, use these.
| What you want to do | Command to run in Terminal |
|---|---|
| Check if OpenClaw is running | openclaw gateway status |
| Restart OpenClaw | openclaw gateway restart |
| Open the dashboard in browser | openclaw dashboard |
| Diagnose problems automatically | openclaw doctor |
| See what's happening live | openclaw logs --follow |
| Check channel connections | openclaw channels status |
Most problems are fixed by running: openclaw doctor
It finds issues and tells you how to fix them.
Full docs: docs.openclaw.ai
$100 in free cloud credits for every attendee tonight.
The ultimate cloud for AI explorers
Build, tune, and run your AI models and applications on top-notch NVIDIA GPUs. Nebius gives you the compute infrastructure to run the kind of agent pipelines you built tonight.
Claim your free credits using the link below:
Claim $100 Nebius Creditnebius.com/promo-code