Seven Figure Agency
Seven Figure Agency Training

MASTER
OPENCLAW

The complete training hub for installing, configuring, and running OpenClaw — your self-hosted AI agent gateway. Built for Seven Figure Agency members who want to move fast.

Start Setup Cold Outreach Guide
350+
Active Member Agencies
187
Members Over $7 Figures
$250M+
Combined Revenue Generated
01 — Understanding OpenClaw

WHAT IS OPENCLAW?

OpenClaw is a self-hosted, open-source AI agent gateway that connects your messaging apps — WhatsApp, Telegram, Discord, iMessage — to powerful AI models. You own the data. You control the agents. You set the rules.

Gateway

The always-on core process that manages routing, control plane, and all channel connections from a single multiplexed port.

Channels

The messaging platforms you connect: WhatsApp, Telegram, Discord, iMessage, and more. Each channel is a live two-way bridge to your AI.

Agents

The AI personalities and configurations that process your messages, execute tasks, and respond intelligently based on your instructions.

Models

The underlying LLMs powering your agents — OpenAI GPT-4, Anthropic Claude, Google Gemini, and others. Switchable on the fly.

Tools & Skills

Capabilities granted to agents: web search, code execution, file management, API calls, and more. Discovered and installed via ClawHub.

ClawHub

The public skill registry for OpenClaw. Browse, install, and manage skills that extend your agent's capabilities with one command.

02 — Installation

INSTALL OPENCLAW

Choose your installation method. The installer script is the fastest path for most users. Hostinger VPS is recommended for always-on production deployments.

Recommended: One-Line Installer

Works on macOS, Linux, and WSL2. Handles Node.js detection, installation, and the initial onboarding wizard automatically.

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

After running, follow the prompts from the openclaw onboard wizard to complete setup.

03 — Configuration

CORE SETUP

After installation, configure your API keys and access the web dashboard. Your Gateway Token is the master key — keep it secure.

OPENCLAW_GATEWAY_TOKENRequired

Your master access token for the web UI. Auto-generated on first install. Never share this.

OPENAI_API_KEYOptional

For GPT-4 and GPT-4o models. Get yours at platform.openai.com.

ANTHROPIC_API_KEYOptional

For Claude 3 Opus, Sonnet, and Haiku models.

GEMINI_API_KEYOptional

For Google Gemini Pro and Flash models.

OpenClaw Setup

Accessing the Web Dashboard

  1. 1Find your VPS IP and the assigned port (default: 18789) in Docker Manager.
  2. 2Navigate to http://<your-vps-ip>:<port> in your browser.
  3. 3Enter your OPENCLAW_GATEWAY_TOKEN when prompted.
  4. 4You're in. The dashboard shows Gateway status, Channels, and Agents.
04 — Channels

CONNECT YOUR CHANNELS

Connect your messaging platforms to OpenClaw. Each channel becomes a live interface to your AI agent.

W

WhatsApp

QR Code Linking
1

Show QR Code

In the OpenClaw web UI, navigate to Channels and click Show QR next to WhatsApp.
2

Link Your Device

Open WhatsApp on your phone → Settings → Linked Devices → Link a Device. Scan the QR code.
3

Troubleshoot status=515

If you see a status=515 error, go to Settings → Config → Update in the UI, then return to Channels.
T

Telegram

BotFather Token
1

Create a Bot

Open Telegram, search for @BotFather, and send the /newbot command.
2

Get Your Token

Follow BotFather's prompts to name your bot. It will provide a bot token — copy it.
3

Connect in OpenClaw

In the OpenClaw web UI, go to Channels → Telegram and paste your bot token into the required field.
05 — Optimization

OPTIMIZE YOUR SETUP

Get more out of OpenClaw with model selection, skills, and operational best practices.

Model Selection

Switch your active model directly from any connected chat interface.

bash
# List available models
/model list

# Switch to a specific model
/model openai/gpt-4.1-mini
/model anthropic/claude-3-5-sonnet
/model google/gemini-2.0-flash
ClawHub Skills

Browse clawhub.ai to discover skills. The API Gateway skill is essential for Google Workspace integrations.

API Gateway
Connect to external APIs and services
Brave Search
Real-time web search capability
Email Validator
Validate email addresses at scale
Code Runner
Execute code in a sandboxed environment
Operations
Gateway Management
bash
openclaw gateway status
openclaw gateway restart
openclaw gateway stop
openclaw logs --follow
openclaw doctor
Security

Never share your OPENCLAW_GATEWAY_TOKEN. Use channels.whatsapp.allowFrom in your config to restrict who can interact with your bot.

Updating

On Hostinger: Docker Manager → Update. Via CLI: npm install -g openclaw@latest then restart the gateway.

06 — Settings & Automation

SETTINGS INSTRUCTIONS & HEARTBEAT

Control your agent's behavior through workspace Markdown files — and automate proactive check-ins with Heartbeat.

Settings Instructions

OpenClaw has no single "settings" panel. Instead, the agent reads a set of Markdown files from its workspace (~/.openclaw/workspace/) on every session turn. Edit these files to control behavior, persona, and memory.

AGENTS.md

Operating rules, priorities, and 'how to behave' instructions. The primary place to write your agent's instructions.

SOUL.md

Persona, tone, and boundaries. Define the agent's character and communication style.

USER.md

Who you are, your business, preferred name, and personal context the agent should always know.

IDENTITY.md

The agent's name, emoji, and vibe. Created/updated during the bootstrap ritual.

HEARTBEAT.md

A short checklist the agent checks during each heartbeat run. Keep it concise to avoid token burn.

MEMORY.md

Optional long-term curated memory. Only load in private sessions.

Token tip: Every workspace file is injected into the context window on every turn. Keep each file under 500 words to control costs and avoid compaction.

Edit files directly from your connected chat:

bash
/edit AGENTS.md
/edit SOUL.md
/edit USER.md
/edit IDENTITY.md

Heartbeat — Proactive Mode

Heartbeat runs a background agent turn on a timer — no message from you needed. The agent checks HEARTBEAT.md and proactively sends you alerts or updates.

Default: ON every 30 minutes. Disable immediately until your setup is tested to avoid unexpected messages and token usage.

Disable heartbeat first:

bash
/config set agents.defaults.heartbeat.every 0m

Re-enable once HEARTBEAT.md is configured:

bash
/config set agents.defaults.heartbeat.every 30m

Restrict to business hours only (in openclaw.json):

json
heartbeat: {
  every: "30m",
  activeHours: { start: "09:00", end: "22:00" }
}
HEARTBEAT_OK

Agent replies this when nothing needs attention — silently suppressed, no notification sent.

activeHours

Restrict heartbeat to business hours to avoid off-hours pings.

model override

Use a cheaper model (e.g. gemini-2.5-flash) for heartbeat runs to reduce API costs.

07 — FAQ

COMMON QUESTIONS

READY TO RUN COLD OUTREACH?

Once OpenClaw is set up, the next step is deploying the cold outreach workflow. 32 meetings in 72 hours. Here's how.

View Cold Outreach Guide