Deploy Ollama + Open WebUI on Your VPS
Deploy Ollama with Open WebUI on your VPS in one click. Run open LLMs — Llama, Mistral, Qwen, Gemma — on your own server, behind a ChatGPT-style interface.
Ollama runs open large language models on your own hardware, and Open WebUI puts a familiar ChatGPT-style interface in front of them. Pull a model from the UI, chat with it, upload documents to ask questions against, and give your team accounts — all without a single token leaving your server or a cent going to an API provider.
Vessl deploys the pair wired together. Ollama has no authentication of its own, so it is kept off the public internet entirely and reachable only over the project network; Open WebUI, which has real logins, is the only public door. The first account you register becomes the admin.
What's included
- Ollama + Open WebUI deployed together and wired up for you
- Pull Llama, Mistral, Qwen, Gemma, and more from inside the UI
- Ollama kept internal-only — no unauthenticated model endpoint exposed
- Multi-user chat with history, document upload (RAG), and prompt library
- Model weights on a persistent volume — no re-downloading on redeploy
Common use cases
- Private ChatGPT for a team, with no data leaving your infrastructure
- Experimenting with open models without per-token API bills
- An OpenAI-compatible local endpoint for your own apps to call
Frequently asked questions
Do I need a GPU?
Not to run it — but be realistic. On a CPU-only VPS, stick to small models (roughly 1–3B parameters) and expect slow replies. For larger models at usable speed, deploy this to a server with a GPU.
Why is Ollama not given its own URL?
Ollama has no authentication at all. Anything that can reach it can run inference, pull arbitrary models, and fill your disk. So it stays on the internal network and Open WebUI, which has logins, is the only thing exposed.
Can my own apps use the models?
Yes — services in the same Vessl project can call Ollama directly over the internal network on port 11434, and Open WebUI also exposes an OpenAI-compatible API for authenticated clients.
Ready to ship?
Deploy Ollama + Open WebUI in under a minute.
Connect your VPS, pick this template, fill in any required fields. Vessl handles the container, SSL, and persistent storage.
Start for FreeNo credit card · BYOS · IDR billing