Pre-Flight Check

Before deploying SuperClaw, you need OpenClaw running. This chapter gets your AI backend online and introduces the SuperClaw CLI — your tool for installing and managing the SuperClaw dashboard.

What is SuperClaw?

SuperClaw is two things working together, sitting on top of OpenClaw.

OpenClaw is the AI backend that does the actual work — conversations, memory, tools, channel integrations. SuperClaw doesn't extend what OpenClaw can do. It gives you a proper interface for controlling and monitoring it.

System Requirements

• Node.js 18+ — runtime for both OpenClaw and the dashboard
• Terminal access — you'll be running CLI commands throughout this guide
• An AI provider account — OpenClaw works with Claude, OpenAI, Gemini, and others. Claude is recommended and used by default

# Verify Node.js version
node --version
# Should show v18.0.0 or higher

Do You Need Memory Embeddings?

Short answer: if you're running multiple agents and more than one person is messaging your AI, yes.

Here's why it matters. Imagine you run a two-person business — you and a partner. You've both connected to OpenClaw via Slack. You're in #dev talking about a bug in your checkout flow. Your partner is in #marketing asking about last quarter's campaign results. A third channel, #support, is fielding customer questions.

Over time, three specialized agents build up their own memory:

This is exactly what SuperClaw is built for. The message router sends "fix the checkout bug" to the right agent automatically, regardless of which of you sent it. Each agent maintains isolated memory so context doesn't bleed between domains.

But here's what happens after a few months: memory gets wide. The developer agent remembers 50 technical decisions. The marketing agent has campaign notes going back to launch. The support agent has patterns from hundreds of customer interactions. And sometimes the context you need sits across all of them — a customer complaint that's actually a billing bug that marketing accidentally mentioned in a campaign brief.

Without embeddings, your AI searches memory by keyword — it finds what you literally said. With embeddings, it finds what you meant. A question like "what was the issue with the payment flow last quarter?" retrieves the right thread even if nobody used those exact words at the time.

Setting Up an Embedding Provider

The easiest free option is Gemini. Grab a key from Google AI Studio in about two minutes, then run:

openclaw auth add --provider gemini

---

Step 1: Install OpenClaw

npm install -g openclaw

Verify it's working:

openclaw --version

Step 2: Configure OpenClaw

Run the interactive setup wizard:

openclaw configure

This walks you through:

• Connecting your AI provider (Anthropic recommended — uses your Claude subscription via OAuth, or paste an API key from console.anthropic.com)
• Configuring your workspace directory
• Setting up any channel integrations (Slack, Telegram, Discord, etc.)

Step 3: Start the Gateway

openclaw gateway start

The Gateway is the core OpenClaw process — it handles AI conversations, message routing, and tool execution. It runs in the background and stays running.

Check it's healthy:

openclaw health

Step 4: Install the SuperClaw CLI

npm install -g @skunkceo/superclaw-cli

Verify it installed:

superclaw version

Step 5: Install the Dashboard

superclaw init

This clones the SuperClaw dashboard from GitHub and sets it up locally. The wizard asks where you want to install it and walks you through creating your first admin account.

Once installed, access the dashboard at the URL shown in the output (typically http://localhost:3000).

---

SuperClaw CLI Reference

---

Free vs Pro Dashboard

---

Time Investment

This chapter: ~15 minutes Full guide: ~2 to 3 hours

---