Onboarding Wizard

What is the Onboarding Wizard?

The Onboarding Wizard is a guided, 5-step setup flow that runs immediately after your server finishes provisioning. Its purpose is to configure the core components your AI agent needs to function: your AI provider credentials, a messaging channel, and your agent's personality. Each step builds on the previous one, so by the time you reach the final step your agent is fully operational and ready to receive messages.

You can skip the Onboarding Wizard entirely and configure everything later from the Agent Control Panel in your dashboard. However, completing the wizard right away ensures your agent is live and responding within minutes of deployment.

Step 1: Confirm server name

The first step confirms the name you chose during the Deploy Wizard. You see your instance name displayed along with the server status (which should show "Online" at this point). If you want to change the name, you can edit it here. This name is for your reference in the dashboard only.

Click Continue to proceed to the API key setup.

Step 2: Set up your API key

This is the most important step. You choose your AI provider (Anthropic, OpenAI, OpenRouter, xAI, Google Gemini, or any of the 20+ supported providers) and then paste your API key. The wizard validates your key in real time by making a lightweight test call to the provider's API.

If the key is valid, you see a green checkmark and a list of available models for that provider. Select the model you want your agent to use. If the key is invalid or expired, the wizard shows an error explaining what went wrong and how to fix it.

Step 3: Connect a channel

In this step, you connect your agent to a messaging platform so people can actually talk to it. OneClickClaw currently supports three channels:

• Telegram — You need a bot token (obtained from BotFather on Telegram) and optionally a group chat ID where the bot should operate. The wizard provides step-by-step instructions for obtaining both values.
• WhatsApp — You need a WhatsApp Business API access token from the Meta Developer dashboard. The wizard guides you through connecting your WhatsApp Business account.
• Discord — You need a Discord bot token from the Discord Developer Portal. The wizard walks you through creating a bot application, generating a token, and inviting the bot to your server.

You can connect one channel during onboarding and add more later from the Agent Control Panel. Your agent can serve multiple Telegram groups and Discord servers simultaneously. For detailed channel setup instructions, see Telegram Setup and Discord Setup.

Step 4: Write your system prompt

The system prompt defines your agent's personality, behavior, and knowledge boundaries. It is sent to the AI model at the start of every conversation and controls how your agent responds. This is where you make your agent unique.

The wizard provides a default system prompt to get you started. You can use it as-is, modify it, or write your own from scratch. A good system prompt typically includes:

• Identity: Who the agent is. "You are a helpful customer support assistant for Acme Corp."
• Tone and style: How it should communicate. "Be friendly, concise, and professional. Use simple language."
• Knowledge boundaries: What it should and should not discuss. "Focus on our products and services. Do not give medical or legal advice."
• Behavior rules: How it handles edge cases. "If you do not know the answer, say so and offer to connect the user with a human."

You can refine your system prompt at any time from the Agent Control Panel. Many users start with the default and iterate over days and weeks as they learn what works best for their audience.

Step 5: Review and confirm

The final step shows a summary of everything you configured: your provider, model, connected channels, and system prompt. Review each section to make sure everything looks correct.

If you connected a channel, the wizard sends a test message through it to verify the entire pipeline is working: your API key, your model selection, and your channel connection. If the agent responds successfully, you see a confirmation with a green "All systems go" message.

If the test fails, the wizard identifies which part of the chain broke (API key issue, model error, or channel connection problem) and offers guidance on how to fix it. You can go back to any previous step and adjust your settings without losing your other selections.

Tips for a smooth setup

The entire Onboarding Wizard typically takes under 3 minutes if you come prepared. Here is how to make it as smooth as possible:

• Have your API key ready before you start. Log into your AI provider's dashboard and copy the key to your clipboard so you can paste it quickly during Step 2.
• Create your Telegram or Discord bot in advance. Setting up a bot on Telegram (via BotFather) or Discord (via the Developer Portal) takes a few minutes. Doing this beforehand means you can breeze through Step 3.
• Check your provider balance. Make sure your AI provider account has credits loaded or an active payment method. The test in Step 5 fails if the provider rejects the API call due to insufficient funds.
• Draft your system prompt. If you already know what personality and behavior you want, write it out in a text document before starting the wizard. You can paste it directly into Step 4.
• You can always come back. If you do not have a channel or system prompt ready, skip those steps and configure them later from your dashboard. The wizard saves your progress so you do not lose your provider and model selections.