Deploying AI Models on a VPS: A Step-by-Step Guide

In the AI era, running models on a Virtual Private Server (VPS) allows for scalable, cost-effective deployments. Whether you're experimenting or deploying AI-powered applications, this guide walks you through setting up your VPS, installing AI models, and running them with both command-line and GUI-based solutions.

Why Choose a VPS for AI?

Hosting AI models on a VPS provides:

• Scalability: Easily upgrade resources as needed.
• Cost Efficiency: Avoid expensive cloud AI services.
• Customization: Full control over installed packages and configurations.
• Accessibility: Run AI models from anywhere.

Let’s get started with setting up a VPS and running AI models.

Step 1: SSH into Your VPS

Once you've purchased a VPS from a provider like AWS, DigitalOcean, or Linode, access it using SSH.

ssh -i "path_to_your_private_key.pem" root@your_server_ip

If prompted, confirm the connection to establish secure access.

Step 2: Update and Upgrade Your System

Before installing any software, ensure your system is up-to-date.

sudo apt-get update && sudo apt-get upgrade -y

This refreshes your package lists and installs any pending security patches or system updates.

Step 3: Installing & Running an AI Model Locally

If you want to run models in the terminal without a GUI, use Ollama, a lightweight model runner.

Installing Ollama

curl -fsSL https://ollama.com/install.sh | sh

Running a Model

ollama run deepseek-r1:7b

For detailed debugging:

ollama run deepseek-r1:7b --verbose

This command downloads and runs the DeepSeek-R1 7B model locally.

Step 4: Running AI Models with a GUI

For easier model management, use Coolify, a self-hosted platform that simplifies AI deployments.

Installing Coolify

curl -fsSL https://cdn.coollabs.io/coolify/install.sh | bash

Accessing Coolify

After installation, navigate to:

http://your_server_ip:8000

Setting Up Coolify

1. Create an Admin Account – Enter your email and password.
2. Optional: Configure a Custom Domain – Set up your domain in settings.
3. Create a New Service – Click on ‘New Service’ and select Ollama GUI.

Adding AI Models

1. Sign in to the Ollama GUI.
2. Click ‘Add Model’ and choose an open-source model.
3. Congratulations! Your AI model is now self-hosted and accessible.

Final Thoughts

Setting up a VPS to run AI models provides full control over deployments, whether for personal use or enterprise solutions. This guide covered:

• SSH access and package updates
• Running AI models locally with Ollama
• Deploying AI models via GUI using Coolify

What’s Next?

In our next blog, we’ll explore advanced optimizations, including:

• Fine-tuning AI models for better accuracy
• Deploying models with GPUs for performance boosts
• Automating deployments with Docker and Kubernetes

Stay ahead in the AI revolution—start deploying your models today!