Build Your First AI Assistant on a VPS in 15 Minutes With OpenClaw
Publicado no IA Dots ·
Agentes de IAAI SignalAssistente pessoaldot-ai_agentsdot-ai_assistantsdot-artificial_intelligenceidioma-enInglêsyt-Ih9OfoGzD54
Vídeo em Inglês, publicado pelo canal AI Academy.
About this content: The summary, key points and topics below are generated automatically from the video transcript. They reflect the content, opinions, claims and predictions presented in the video and do not represent a position, validation or fact-check by IA Dots.
Video summary
This video provides a step-by-step beginner-friendly guide on how to build and run an AI assistant using Open Claw on a virtual private server (VPS). It covers choosing the right hosting provider, installing the software, connecting chat apps, configuring operational briefs, and implementing safe workflows and permissions.
Key points
- A virtual private server (VPS) allows Open Claw to run 24/7 online independently of a personal laptop.
- Open Claw requires specific minimum system specifications, such as at least 4 GB of RAM and a Linux environment like Ubuntu.
- Managed hosting providers simplify the deployment process, often taking only a few minutes to get the dashboard live.
- Connecting a chat app like Telegram or WhatsApp enables interaction with the AI assistant directly from daily messaging platforms.
- Configuring a personalized brief helps avoid generic responses by giving the AI context about the user's role, priorities, and tone.
- Establishing strict permissions and starting with limited access prevents the assistant from executing unauthorized actions like sending messages or making purchases.
Topics
AI AssistantOpen ClawVPS SetupHostingAutomationChat AppsPermissionsWorkflows
What you will learn
- Open Claw requires a dedicated VPS with root access and a minimum of 4 GB of RAM, meaning shared hosting plans will not work.
- Using a separate phone number for WhatsApp integration prevents the AI assistant from accessing private personal conversations.
- Providing a specific brief containing context, constraints, and tasks transforms the AI from a generic chatbot into a functional co-worker.
- Limiting initial permissions and utilizing approval requirements for actions like sending messages or deleting files protects against errors.
Questions and answers
What are the minimum system requirements for running Open Claw on a VPS?
Open Claw requires at least 4 GB of RAM (8 GB if using browser automation), NVMe or SSD storage, stable uptime, and a Linux environment such as Ubuntu 22.04 or 24.04.
Why is shared hosting not recommended for Open Claw?
Shared hosting does not support background processes, Docker containers, and persistent resources which are necessary for Open Claw to function.
How can users prevent Open Claw from giving generic responses?
Users can fix this by creating a simple brief in the dashboard chat that explains who they are, their priorities, preferred tone, and what actions always require approval.
Video transcript
Transcript from available captions; it may contain errors.
Most people hear VPS and close the tab. Server room, command line, headache. No, thanks. Here's what they don't know. With the right provider, you can have Open Claw running on your own private server online 24/7 for about $6.99 a month. And the whole setup takes less time than deciding what to watch next. What's up, guys? My name is Kyle from AI Academy. And in this video, I'm walking you through how to build your first AI assistant using Open Claw on a VPS. Step-by-step beginner-friendly no technical background needed. By the end of this video, you will have Open Claw installed, connected to a chat app, and ready to start handling real tasks for you. Running 24/7 on your own private server. This can save you a lot of time and frustration online. So, let's get into it. Now, before we get into the setup, let me quickly explain what a VPS is. Because this is the part that scares most beginners off, and honestly, it shouldn't. VPS stands for virtual private server. Think of it like renting a small computer that lives in a data center and stays online all the time. Even when your laptop is closed, when you're asleep, even while you're on holiday. It's always running, always available. Now, here's why that matters for Open Claw specifically. Open Claw is a self-hosted AI assistant. That means it doesn't live on someone else's platform. It lives on your server. Your data stays private, and your assistant stays online around the clock. And nobody can shut it down, change the pricing, or change the rules on you. The alternative is running Open Claw on your own laptop, which means the moment you close the lid, your AI assistant goes offline. A VPS solves that completely. And here's the thing most beginners don't realize. You don't need to be a developer to set this up. You don't need to know what a command line is. With the right hosting provider, this is set up and running in minutes. And that's exactly what we're doing today. So, let's get into step one. Step one is picking your VPS. And this decision matters more than most tutorials let on. Open Claw has minimum requirements that it needs for it to run properly. At least 4 gig of RAM, 8 GB of RAM if you plan to use browser automation features, NVMe or SSD storage, a Linux environment, Ubuntu 22.04 or 24.04 works best, stable uptime, and consistent network connection. Here's the beginner mistake almost everyone makes. They grab the cheapest shared hosting plan they can find and wonder why Open Claw isn't installing. Shared hosting doesn't work. Open Claw needs background processes, Docker containers, and persistent resources, none of which shared hosting supports. You need a VPS with root access. Hostinger managed Open Claw is the easiest starting point. It handles the server environment, security updates, and backups automatically. So, you're not managing any of that yourself. It starts around $6.99 a month on a longer plan, and it includes AI credits preloaded, so you don't need to set up external API accounts separately. If you want full control and you're comfortable with a bit more setup, a standard VPS on Hostinger, Digital Ocean, or Hetzner works great, too. Just make sure your plan meets the RAM requirements. Step two, install Open Claw on your VPS. Go ahead and click the link in the description below and get your hosting account set up so you can follow along with me. Your screen should look exactly like mine. If you're using Hostinger managed Open Claw option, this step is almost automatic. Here's what you need to do. Log in to your Hostinger dashboard, in the left column, select your Open Claw account, and then you will see deploy. Click deploy, and it will then deploy. If you're setting up Open Claw manually on a standard VPS, the process takes a bit longer, but it's still straightforward. Log in to your account. In the left column, select VPS. You can then select manage, Docker manager, and then select compose. Then select one-click deploy, then type in OpenClaw, then select, scroll down, and then click deploy. Make sure to choose your AI model and provider during setup. Either way, whichever one you use, you will have access to OpenClaw's dashboard. That's your control center for everything going forward. One thing to note here, during onboarding, OpenClaw can detect existing CodeX CLI or Claude code access if you already have the setup. If you do, it'll automatically configure itself for you to use it. If you don't, no problem. Just follow the standard provider setup and you're good. OpenClaw is installed and your dashboard is live. That's the hard part done, and honestly, it wasn't hard at all. But here's where most people get stuck next. They connect their chat app, send a test message, and the response feels completely generic, like talking to any other chatbot. The fix is what we're covering in step three, and it's the difference between an AI assistant and an AI co-worker. Now connect OpenClaw to the chat app that you use every day. This is where OpenClaw starts feeling real. To do this, in the left column, select main and then go to agent settings. Scroll down to channels. Here you will have a list of all your channels. As you can see, I have Telegram connected. If you would like to connect WhatsApp, Discord, or another chat app, simply go to it and then select setup and then follow the prompts. Once you have set that up, send a test message saying, "Hello, what can you help me with today?" If you get a response, as I did, that means it is live. If you're connecting WhatsApp specifically, and this is important, do not link your personal number, set up a separate number. Specifically dedicated to your OpenClaw assistant. Your personal WhatsApp has private conversations with family members and sensitive information. You don't want your AI assistant getting a hold of all of those messages. Use a separate number, keep it clean from day one. Now, make sure your account is still open from the link in the description below so you can follow along with me. Step four, this is the most that tutorials skip and it's the most important one. A fresh OpenClaw installation doesn't know anything about you. It doesn't know what you do, how you like things written, what it can handle on its own, what it needs approval for. Without this configuration, every response will feel generic. Fix that by creating a simple brief. So, in your dashboard's chat, you want to tell OpenClaw a few things. You want to tell it who you are and what you do. This could be one or two sentences. You can say you own a small e-commerce business selling handmade goods. I handle everything from customer communication to orders. Your priorities, what matters most right now, client emails, content creation, research, tell it. Your preferred tone, professional casual friendly or direct. Make sure to be specific. What always needs your approval, sending messages, making purchases, deleting files, accessing private data. These should always require a confirmation from you, always. Here's an example of what that last rule looks like in practice. Draft emails freely but never send anything without my approval first. That one line prevents a huge number of problems. Step five, put OpenClaw to work and do it safely. Start with three beginner workflows that start well right out of the gate. Every morning, message OpenClaw and tell it this. Here is what's on my plate today, help me pick my three most important tasks. Then paste in your tasks, then let OpenClaw help you prioritize. Secondly, a writing partner. You can send it rough notes and tell it to turn these bullet points into a professional client update. Keep it under 150 words. Give it the source material, the format, the tone. This is the formula. Raw inputs plus clear instructions equals something you can actually use. Next, you could also use it as a research assistant. You could ask it to compare these three tools for a small business. Site sources and recommend one. Give it context about your budget and use case. It'll put the comparison together and give you a reasoned recommendation in minutes instead of hours. Now, on permissions, that part matters. Open Claw can run tools, access files, and send messages if you allow it to. That's powerful. It's also why you need to be careful about what you turn on, especially early on. The rules are simple. Start with limited access. Watch what it does before you expand what it can do. Use an allow list to control who can message your chatbot. Keep background automations disabled until you understand exactly what they're doing. Require your approval for anything related to messages being sent, files deleted, purchases made, or private data touched. Here are five quick mistakes that trip up almost every beginner. Mistake one, connecting every account on day one. One channel, one workflow, that's your starting point. Connecting everything at once leads to overwhelm and nothing working properly. Mistake two, sending vague prompts. Help me with this is not a task. Use this formula every time. Context, task, constraints, output formula. The more specific you are, the better the response. Mistake three, expecting perfect memory. Open Claw doesn't automatically carry context between conversations. Give it context each time until that is set up properly. Mistake four, granting permission too early. We just covered this, but it's worth repeating because it's one that causes real headaches. Review before it acts. And mistake number five, treating it as a decision-maker. Open Claw is a force multiplier. It makes you faster and more organized. It doesn't replace your judgment. So, make sure you keep yourself in the loop. Avoid these five steps and you're already ahead of the game. And that's it. Your first AI assistant on a VPS. Quick recap. Choose the right VPS. 4 gig RAM minimum, SSD, Ubuntu root access. Install Open Claw. Manage hosting makes this 2 minutes. Manual setup gives you full control. Connect one chatbot. Telegram for most beginners. And make sure you have a separate number for WhatsApp. Configure Open Claw with proper briefing before you throw tasks at it. And lastly, start with three workflows. Planning, writing, research, and keep permissions tight. Now, if you haven't set up a VPS, you can set one up by using the link in the description below. No technical knowledge required. And if this video helped, make sure to smash that like button, subscribe, and hit that bell icon. My name is Kyle from AI Academy. Thank you so much for watching, and I'll see you in the next one.
Assistir ao vídeo original no YouTube
Assuntos: Agentes de IA, Assistente pessoal, Assistente pessoal. Publicado no YouTube em 2026-10-15.
Seleção automática por assunto e idioma. O vídeo pertence ao canal de origem.
About OpenClaw
OpenClaw is an open-source AI assistant that works as an autonomous agent. It can browse the web, fill in forms, read and write files and run system commands, with full access or in a sandboxed environment. It is used through messaging apps such as WhatsApp, Telegram, Discord, Slack, Signal and iMessage, keeps memory between conversations and can be extended with skills and plugins. It works with hosted or local language models, and the data stays on the user's machine.
The project was created by Peter Steinberger and was previously called Clawdbot and Moltbot. It is written in TypeScript, released under the MIT license and managed by the OpenClaw Foundation, a US nonprofit organization (501(c)(3)).
Fact sheet
- Software: OpenClaw (formerly Clawdbot and Moltbot)
- Type: Open-source AI assistant and autonomous agent
- Creator: Peter Steinberger
- Managed by: OpenClaw Foundation (nonprofit)
- Language: TypeScript
- License: MIT
- Operating systems: macOS, Windows, Linux
- Repository: github.com/openclaw/openclaw
- Official website: openclaw.ai
Read also
- How to Use OpenClaw as Your Personal AI Coworker (Beginner Guide)
- Mac Mini 16GB vs 32GB for AI Agents: OpenClaw Uses 1GB
- OpenAI, Meta & Claude All Launched AI Agents This Week. We Built Ours Today.
Sources and references: Wikidata (Q137916205) · official OpenClaw website · GitHub repository · video by AI Academy on YouTube.