Full OpenClaw Setup Tutorial: Step-by-Step Walkthrough (Clawdbot)
Introduction to OpenClaw
Overview of OpenClaw
- Caleb introduces OpenClaw as a unique AI assistant that operates within existing applications, unlike traditional tools like ChatGPT.
- The assistant can run tasks in the background on a schedule, reducing the need for constant user interaction.
Setting Up OpenClaw
- Any machine can be used to run OpenClaw, but it must remain online continuously for effective operation.
- Running OpenClaw on personal devices poses security risks and requires maintaining an always-on setup.
Choosing the Right Environment
Importance of VPS
- A Virtual Private Server (VPS) is recommended for running OpenClaw due to its reliability and accessibility from anywhere.
- Hostinger is suggested as a cost-effective VPS option with beginner-friendly deployment features.
VPS Setup Process
- Users are guided through selecting a VPS plan with sufficient resources (e.g., KVM 2 plan).
- The video explains billing options, recommending at least a 12-month plan for savings and benefits like a free domain.
Finalizing Your VPS Configuration
Completing Registration
- Instructions are provided for registering an account with Hostinger and entering payment details.
- Users are advised to save their OpenClaw gateway token securely, treating it like a password.
API Key Requirements
- The setup requires API keys from both OpenAI and Anthropic; users must ensure billing is enabled on these platforms.
- While using OpenClaw is free, costs arise from utilizing AI models based on usage metrics.
OpenClaw Setup and Integration
Creating API Keys for OpenClaw
- The project is named "OpenClaw," and users are instructed to set their project permissions before creating a secret key.
- Users can opt to set up both OpenAI and Anthropic APIs, allowing flexibility in model selection based on cost or suitability.
- After filling out the API keys, users must click "Deploy" to initiate the deployment process, which may take several minutes.
Connecting to the OpenClaw Dashboard
- Once deployed, users will see a status change from "Created" to "Running." They need to connect by pasting the gateway token into the dashboard.
- A test message is sent through the chat feature to confirm that everything is functioning correctly.
Setting Up Telegram Integration
- The video transitions into setting up a chatting app like Telegram, highlighting its popularity due to easy bot setup and token connections.
- Users are guided to send a message requesting steps for setting up Telegram integration with their assistant.
Creating a Bot with BotFather
- Instructions include accessing BotFather via a provided link and initiating bot creation by typing "/newbot."
- Users must name their bot (e.g., "OpenClaw Tutorial 123") and choose a unique username ending with "bot."
Configuring Chat ID and Finalizing Setup
- After obtaining the bot token from BotFather, it needs to be sent as a message in OpenClaw's dashboard.
- To get the chat ID, users interact with another bot linked in the description. This ID is then pasted back into OpenClaw for configuration completion.
Testing Telegram Functionality
- Users are prompted to send messages through Telegram after configuration. Successful communication confirms that everything is working properly.
Model Confirmation and Skills Installation
- Users check which AI model (Claude Opus 4.5 or GPT-4o) is currently active; they can request changes through their assistant if needed.
- The video introduces ClawHub as an option for installing additional skills that enhance assistant capabilities.
How to Install and Manage Skills for Your AI Assistant
Installing Skills from GitHub
- Users can publish and browse skills through the public skills directory on GitHub, with installation instructions provided in the video description.
- The Google Workspace skill is highlighted as a useful starting point, allowing integration with Gmail, Calendar, Drive, Contacts, and Sheets for enhanced AI assistant functionality.
- To install a skill, copy its GitHub link and send a message to your bot in Telegram stating "install this skill," followed by the link. The bot will guide you through any necessary troubleshooting steps.
Handling Dependencies
- If an error occurs during installation (e.g., "brew not installed"), it indicates that additional dependencies are required. These can be managed via the Docker manager on Hostinger.
- Users should communicate directly with their VPS using a built-in terminal to install missing dependencies and refresh the setup.
Setting Up Recurring Actions
- Cron Jobs allow users to automate tasks for their AI assistant. Access this feature through the OpenClaw control panel under Cron jobs.
- Create a job by naming it (e.g., "Summarise Emails and Meetings") and setting a schedule (e.g., daily). Define what actions the AI should perform within the system text box.
Managing Deployments on Hostinger
- The Hostinger Docker manager serves as a control panel for managing deployments of OpenClaw, including restarting or redeploying services and changing API keys or environment variables.
- Users can configure various settings such as OS changes, backups, monitoring setups, security configurations, or domain connections.
Security Best Practices
- Given that OpenClaw is experimental and has access to sensitive data like emails and calendars, it's crucial to prioritize security measures post-installation.
- Enable automatic backups in Hostinger's hPanel; upgrading from weekly to daily backups is recommended for quicker recovery options.
- Secure dashboard access by using strong passwords and limiting access based on IP addresses or VPN usage.
- Consider implementing SSL certificates when adding domains for encrypted connections to dashboards.
- Treat tokens (gateway token, Telegram bot token, API keys) like passwords; avoid sharing them publicly. Change any leaked tokens immediately within your Hostinger environment variables.