Ready-to-use VPS
Ollama VPS: your own private ChatGPT-style AI server
Chat with open-source models on a server only you control. We install Ollama and the Open WebUI chat interface on first boot, pull a starter model and create your admin login — open the browser and start chatting.
Why run it on our VPS
Ollama + Open WebUI
The popular local-model runtime with a polished ChatGPT-style web interface on port 8080.
Starter model included
A small CPU-friendly model (Qwen3) is pulled on first boot; add Llama, Gemma, Mistral and more in one click.
Admin login ready
Your admin account is created automatically and public sign-up is closed — no race for the first account.
Private by design
Prompts and documents stay on your VPS. The Ollama API listens on localhost only.
No per-token fees
Pay only for the VPS — unlimited chats with the models you host.
Hourly or prepaid
Hourly billing with a monthly cap, or prepaid Saver plans for the lowest fixed price.
Recommended plans
Prepaid (Saver) — lowest price
Hosting provider or agency? Partners get 30% off hourly VPS.
How it works
- Pick a plan below (4 GB RAM minimum, 8 GB recommended) and sign in or create an account.
- Ollama + Open WebUI is preselected — choose a hostname and create the VPS.
- After a few minutes open http://YOUR-IP:8080 and log in with the credentials in /root/ollama-info.txt.
Ollama and Open WebUI are free, open-source software. Speed depends on the model size and your plan’s vCPU and RAM.
What people build with it
- A private ChatGPT alternative for you or your team
- Chat with your own documents (RAG) in Open WebUI
- A local model API for n8n, Dify or your own apps
- Testing open-source models before production
Frequently asked questions
Does this VPS have a GPU?
No — models run on the CPU. Small models of 1–4B parameters answer comfortably; 7–8B models work on 16 GB RAM but are slower.
Which model is installed?
Qwen3 1.7B on plans under 8 GB RAM and Qwen3 4B on 8 GB or more. You can pull any model from the Ollama library afterwards.
How much RAM do I need?
4 GB is the minimum for a small model and Open WebUI. 8 GB lets you run 3–4B models smoothly; choose 16 GB for 7–8B models.
Is the Ollama API exposed to the internet?
No. Ollama listens on 127.0.0.1:11434 and Open WebUI talks to it locally. Expose it yourself only behind authentication.
Can other people use it?
Yes. As admin you can invite users or enable sign-up in Open WebUI and control which models they see.
