Skip to main content
50% off all plans, limited time. Starting at $2.48/mo
Ollama

Ollama

Ollama runs LLMs locally on your VPS. Pull and serve Llama 3, Mistral, Phi, Gemma, and Qwen with one command. Self-hosted AI without an API key, REST API compatible with OpenAI clients. The simplest path to private inference, used by 200,000+ developers.

At a glance

175k

GitHub stars

37

Page views

324

Active installs

License MIT Version Latest OS Ubuntu Server 24.04 LTS Min RAM 8 GB IP IPV4,IPV6

Active installs is sample data (staging preview); real metric wires in before launch.

Installed Components

  1. Ollama → a lightweight, extensible framework for running LLMs locally.
  2. Open WebUI → web interface for interacting with Ollama.

Installation Directory

The Docker Compose file for Open WebUI is located at:

/root/open-webui/docker-compose.yml

Working directory for the systemd service: /root/open-webui

Access Information

Web Interface:

  • Open WebUI: http://<server-ip>:8080
  • It may take a few minutes for the UI to start.
  • Access early to complete initial setup and create an admin user.

Ollama API:

  • Base URL: http://127.0.0.1:11434
  • Accessible locally inside the server.

Ollama Usage

Run these commands directly in the terminal:

# Pull a model
ollama pull <model_name>

# Run a model
ollama run <model_name>

Example:

ollama pull llama3.1
ollama run llama3.1

Service Management

# Open WebUI
systemctl status open-webui
systemctl restart open-webui

# Ollama
systemctl status ollama
systemctl restart ollama

More in Data Science

Related apps.

Deploy Ollama now. From $2.48/mo.