Back to blog

Getting Started With RunClaw

Published on
Written byRunClaw Team
Getting Started With RunClaw

RunClaw is the easiest way to run OpenClaw on your own server.

Instead of buying a VPS, installing Docker, pulling the OpenClaw image, wiring up chat channels, and debugging config files, you stay in the browser. RunClaw provisions the server, installs OpenClaw, and connects your bot in one guided flow.

This guide walks you from zero to a live, 24/7 OpenClaw agent.

Before you start

You’ll need:

  • A RunClaw account – you can sign in with Google during the flow.
  • A bot token – Telegram via @BotFather or Discord via the Discord Developer Portal.
  • A payment method – to start your subscription; we include monthly AI credits so you can get going quickly.
  • (Optional) Your OpenRouter API key – only if you prefer to bring your own LLM key instead of using the included credits.

Step 1 – Choose your model

From the homepage, use the Deploy an agent wizard or go straight to the Get Started page.

Pick the model that will power your OpenClaw agent. RunClaw currently supports multiple OpenRouter-backed models, including:

  • Claude 4.5 Sonnet / Opus – great general-purpose and coding performance.
  • GPT-5 – advanced multimodal capabilities.
  • Gemini 3 Flash – fast and cost-effective for quick tasks.

You can change this choice later from your dashboard. Under the hood, RunClaw translates your pick into an openrouter/... model ID and writes it into the OpenClaw config on your VPS.

If you toggle “Use your own API key”, paste your OpenRouter key. RunClaw will pass it through securely so OpenClaw can bill usage to your own OpenRouter account.

Step 2 – Choose where you chat

Next, choose your integration. Today, RunClaw supports:

  • Telegram (fully wired and ready).
  • Discord (fully wired and ready).
  • WhatsApp (coming soon).

Select Telegram or Discord as your channel. This is where you’ll talk to your agent day to day.

Step 3 – Create your channel bot credentials

If you selected Telegram and you don’t already have a bot token:

  1. Open @BotFather in Telegram.
  2. Send /newbot and follow the prompts to name your bot and set a username.
  3. BotFather will reply with a token that looks like: 123456:ABC-DEF1234ghIkl-zyx57W2v1u123ew11
  4. Copy this token.

Back in RunClaw:

  • Paste the token into the Telegram Bot Token field in the wizard or on the Get Started page.

If you selected Discord:

  1. Open the Discord Developer Portal.
  2. Create an application and bot user.
  3. Enable Message Content Intent in Bot settings.
  4. Copy the bot token into the Discord Bot Token field.

RunClaw stores this securely and uses it to configure the selected OpenClaw channel on your server.

Step 4 – Sign in and deploy

When you reach Step 4 of the Get Started flow:

  1. Sign in with Google (if you’re not already signed in).
  2. Review the Configuration Summary:
    • Selected model.
    • Integration (Telegram or Discord).
    • Your account email.
  3. Start your subscription:
    • If you’re new, you’ll be redirected to Stripe Checkout to confirm your plan.
    • Returning users with an active subscription go straight to deployment.
  4. Click Deploy Agent.

From here, RunClaw:

  • Spins up a new VPS for you.
  • Installs Docker and Docker Compose.
  • Pulls the latest OpenClaw image.
  • Writes OpenClaw config with:
    • Your OpenRouter API key (either shared or BYOK).
    • Your selected channel bot token.
    • Default model and assistant settings.
    • dmPolicy: open and allowFrom: ["*"] so you can DM the bot right away.
  • Runs a setup script that:
    • Calls openclaw doctor --fix to ensure channels are correctly enabled.
    • Applies your model and selected channel settings via the OpenClaw CLI.
    • Restarts the container so changes take effect.

You don’t have to SSH into the server or touch a single YAML file.

Step 5 – Talk to your agent

Once provisioning finishes (usually within ~15 minutes):

  • Open your selected channel and send a message to your bot.
  • Your OpenClaw agent is now running 24/7 on your VPS via RunClaw.

You can also open the Dashboard → Chat tab to see a simple in-app chat UI. For now, this is mostly a visual companion—the real-time conversation happens in your selected channel.

Managing your agent from the dashboard

After your first deployment, the dashboard gives you basic controls:

  • Settings
    • See your instance status, server name, and IP address.
    • Send stop, reboot, or delete signals to your instance.
  • Usage
    • View your remaining AI credit balance.
    • See daily and monthly spend.
    • Add funds via Stripe-powered top-ups.
  • Billing
    • Check your subscription status and renewal date.
    • Cancel or resume your subscription.
    • View past invoices and download PDFs.

Troubleshooting basics

If something doesn’t look right:

  • No messages from the bot:
    • Double-check that your channel bot token is correct.
    • Make sure the bot was invited/added to the place you are messaging from.
  • Instance stuck in “pending” or “provisioning”:
    • Wait a few more minutes; cloud provisioning and Docker pulls can take a bit.
    • If it doesn’t resolve, contact support with your account email so we can inspect the status on the server.
  • Usage not updating:
    • Usage data is fetched from /api/usage; try a hard refresh of the dashboard.

If you’re blocked or confused at any point, reach out—we built RunClaw to remove the painful parts of running OpenClaw so you can focus on what your agent actually does.