# Run Your Own AI Employee 24/7 with OpenClaw (100% Local) Source: https://www.youtube.com/watch?v=yuPRgWnPFfA Recap page: https://rapidrecap.app/video/yuPRgWnPFfA Generated: 2026-02-22T09:33:34.499+00:00 --- ## Quick Overview The video successfully demonstrates how to set up an OpenClaw AI assistant to run 24/7 on a remote server using a no-code approach via HostAI.Bot, requiring five main steps: creating a user, installing packages, installing OpenClaw, installing Ollama, and connecting the Telegram channel. **Key Points:** - The installation process for a 24/7 OpenClaw AI employee is achieved in five steps: Create User, Install Required Packages, Install OpenClaw, Install Ollama, and Connect. - The user connects to their remote server (a Hostinger VPS running Ubuntu 24.04) via SSH and performs initial setup, including creating an 'openclaw' user and setting its password (0:08-03:16). - Required packages (Node.js, build-essential, python3) are installed using curl and apt-get, followed by installing OpenClaw via npm (3:17-3:47). - Ollama is installed using its provided script, and the 'ministral-3:3b' model is pulled and configured with OpenClaw (4:04-07:17). - The OpenClaw service ('clawdbot-gateway.service') is created as a systemd service to ensure it runs constantly and restarts automatically upon computer reboot (3:35-04:29). - The final connection step involves using the HostAI.Bot platform to connect Telegram by getting a token from BotFather and approving the pairing code in the HostAI.Bot dashboard, resulting in a live, 24/7 functional AI bot (08:08-11:26). - The setup is presented as a no-code solution, emphasizing ease of use for beginners, even when running on a remote cloud server. ![Screenshot at 00:08: The presenter outlines the three main advantages of running OpenClaw 24/7: Local Computer, Cloud Server, or No Code setup, setting the stage for the remote server tutorial.](https://ss.rapidrecap.app/screens/yuPRgWnPFfA/00-00-08.jpg) **Context:** This tutorial guides viewers through setting up OpenClaw, described as "The AI that actually does things," to run continuously (24/7) on a remote server, offering an alternative to running it locally. The demonstration uses a Hostinger VPS running Ubuntu 24.04 and leverages the HostAI.Bot platform to streamline the complex setup process, which typically involves installing Ollama for local model execution and configuring systemd services. ## Detailed Analysis The video details a comprehensive, no-code tutorial for setting up the OpenClaw AI assistant to run 24/7 on a remote server, primarily using HostAI.Bot for deployment. The process begins by setting up the remote Ubuntu VPS (0:08), starting with user creation for 'openclaw' and granting it sudo privileges without a password (1:58). Next, necessary packages like Node.js and Python 3 are installed (3:18). OpenClaw is then installed globally using npm (3:36), and a systemd service file is created to manage the 'clawdbot-gateway' so it runs constantly and restarts automatically upon system reboot (3:50). Following this, Ollama is installed (4:04), and the 'ministral-3:3b' model is pulled and configured within OpenClaw (6:42). The final stage involves connecting the Telegram channel via HostAI.Bot: the user obtains a token from BotFather (5:37), enters it into the HostAI.Bot dashboard (6:13), approves the channel pairing request (11:55), and launches the bot, which confirms it is live and running on Telegram (11:26). The video concludes by showing the bot interacting successfully via Telegram and the HostAI.Bot dashboard, confirming the 24/7 availability achieved through cloud hosting and automation. ### 24/7 OpenClaw Options - Local Computer - Cloud Server - No Code ### Server Setup (Steps 1-3) - Create User ('openclaw') with sudo access - Install required packages (Node.js, Python3) - Install OpenClaw via npm and create a systemd service for automatic restarts ### Model Installation (Step 4a) - Install Ollama and pull the 'ministral-3:3b' model, configuring OpenClaw to use it (4:04-07:17) ### Getting Dashboard URL (Step 4b) - Execute 'openclaw dashboard' command locally to get the connection URL (7:22-07:42) ### SSH Tunnel (Step 5) - Establish an SSH tunnel from the local machine to the remote server's IP address to access the dashboard locally (08:03-08:23) ### HostAI.Bot Deployment - Sign in with Google, select Telegram channel, use the token from BotFather to pair the bot, and approve the channel pairing request on the dashboard (10:55-12:03) ### Final Status - Bot confirms it is online and running on Telegram, accessible via the HostAI.Bot control panel (11:26-11:32) ![Screenshot at 00:08: The presenter introduces the three primary deployment options for 24/7 OpenClaw: local, cloud server, or no-code managed service.](https://ss.rapidrecap.app/screens/yuPRgWnPFfA/00-00-08.jpg) ![Screenshot at 00:08: The presenter shows the terminal session on the remote server, having just connected via SSH, initiating the command-line setup.](https://ss.rapidrecap.app/screens/yuPRgWnPFfA/00-00-08.jpg) ![Screenshot at 02:54: After logging into the remote server as root, the command line shows the prompt change, indicating successful root access.](https://ss.rapidrecap.app/screens/yuPRgWnPFfA/00-02-54.jpg) ![Screenshot at 03:05: Execution of the command block to create the 'openclaw' user and grant passwordless sudo privileges, although the system reports the user already exists.](https://ss.rapidrecap.app/screens/yuPRgWnPFfA/00-03-05.jpg) ![Screenshot at 05:16: The OpenClaw configuration wizard asks the user to select the default model, with 'openai/gpt-4o-mini' currently selected by default in this session context.](https://ss.rapidrecap.app/screens/yuPRgWnPFfA/00-05-16.jpg)