Running Ollama on a VPS: Step by Step
Running AI models locally is becoming more popular, especially with tools like Ollama. Instead of relying on cloud APIs, you can run powerful AI models directly on your own server.
If you have a VPS, you can use it to host Ollama and run models like Llama, Mistral, or other open-source AI models.
In this guide, we will walk you through running Ollama on a VPS step by step, including installation, configuration, and optimization.
What is Ollama?
Ollama is a tool that allows you to run large language models locally.
It simplifies:
- Model downloads
- Model execution
- API access
Why Run Ollama on a VPS?
Benefits:
1. Full Control
You control data and privacy.
2. No API Costs
No per-request charges.
3. Custom Models
Run your own models.
VPS Requirements
Minimum:
- 4GB RAM
- 2 CPU cores
Recommended:
- 8GB+ RAM
- GPU support
Step 1: Connect to VPS
ssh user@your-server-ip
Step 2: Update Server
sudo apt update && sudo apt upgrade -y
Step 3: Install Ollama
curl -fsSL https://ollama.com/install.sh | sh
Step 4: Verify Installation
ollama --version
Step 5: Run First Model
ollama run llama2
Step 6: Access Ollama API
Default:
- Port 11434
Test:
curl http://localhost:11434
Step 7: Run in Background
Use:
ollama serve
Step 8: Open Firewall Port
sudo ufw allow 11434
Step 9: Access Remotely
Visit:
http://your-ip:11434
Step 10: Optimize Performance
- Use smaller models
- Increase RAM
- Enable swap
Common Issues
Out of Memory
Use smaller model.
Slow Performance
Upgrade VPS.
Best Practices
- Monitor usage
- Secure API
- Use reverse proxy
FAQs
Can I run Ollama without GPU?
Yes.
Is it free?
Yes.
Final Thoughts
Running Ollama on a VPS is a powerful way to host your own AI models. It gives you full control, better privacy, and no dependency on external APIs.