Self-Host Ollama on Your VPS: Run AI Models with One Click

Running AI models locally means no API costs, no rate limits, and complete data privacy. But setting up Ollama on a VPS typically involves SSH, Docker commands, port configuration, and figuring out how to expose the API securely.

Server Compass v1.16.0 turns this into a one-click operation. Deploy Ollama from the Stack Wizard, manage models through a visual interface, and optionally bundle Open WebUI for a full ChatGPT-like experience—all on your own server.

One-Click Ollama Deployment

Select the Ollama template in the Stack Wizard, click Deploy, and Server Compass handles the Docker configuration, port mapping, and GPU passthrough (if available). Your Ollama instance is running in under a minute.

Model Management Tab

After deployment, a new Ollama tab appears in your app dashboard with full model management:

• Browse models — search the Ollama library for Llama 3, CodeLlama, Mistral, Phi, Gemma, and hundreds more
• Install with one click — download and configure models without CLI commands
• Monitor running models — see memory usage per model to manage your VPS resources
• API endpoint access — copy the API URL for use in your applications
• Integrated chat interface — test models directly in Server Compass before integrating them

Ollama + Open WebUI Bundle

Want a polished chat interface for your team? Deploy the Ollama + Open WebUI bundle. Open WebUI provides a ChatGPT-like experience with conversation history, model switching, and multi-user support—all running on your VPS.

5 New CMS Templates

Also in this release: one-click templates for Concrete CMS, Grav (flat-file CMS), MyBB (forum), OctoberCMS (Laravel-based), and phpMyAdmin.

Password Generator

The new password generator icon appears next to password fields throughout Server Compass. Generate secure passwords from 8 to 128 characters with configurable character types—no more using "password123" for your staging databases.

Run AI on Your Own Server

Self-hosting AI models gives you unlimited inference with zero API costs. Try Server Compass and deploy Ollama in one click.