Ollama VPS hosting.

You get a private VPS on which Ollama serves the language models and Open WebUI provides the browser chat interface. Root access, NVMe storage, and DDoS protection come with the server, so you decide which models run, who may log in, and which other self-hosted tools sit next to the stack.

• Private VPS for the full stack
• Ollama runtime plus Open WebUI chat
• Room for storage, proxies, and add-ons

Yes. Order the VPS, open the Coolify-powered One-Click Apps marketplace, pick the Ollama with Open WebUI stack, and deploy it. Containers, persistent volumes, and the connection between Ollama and Open WebUI come prepared, so you skip writing Docker Compose files by hand. Learn more in the Help Center

• Order the VPS first
• Choose the marketplace stack
• Launch from a prepared flow

Yes. Every VPS comes with full root and SSH access. Pull or remove models with the Ollama CLI, change Docker settings, tighten firewall rules, mount extra storage, or update containers on your own schedule — nothing is locked behind a managed layer. How to connect via SSH

• Root and SSH access
• Ollama CLI and Docker control
• Own firewall and storage rules

No GPU is required. Ollama runs models on the CPU, which works well for smaller and quantized models, private testing, internal tools, and lighter always-on workloads. Expect slower token output than on a GPU; when response time or concurrent use becomes the bottleneck, choose more CPU cores and RAM or move to a dedicated server.

• CPU-only works for smaller models
• Good for testing and internal tools
• Scale up when throughput limits you

Yes. Start with a lighter plan for small 1B–3B models, then add RAM, CPU cores, or NVMe storage as users and models grow. Bigger models need considerably more RAM, so size the VPS to the models you want to run. Plan extra headroom for Open WebUI, document embeddings, and several people chatting at once.

• Start small with 1B–3B models
• Larger models (7B, 13B and up) need more RAM
• Add RAM and CPU as usage grows

Yes. Point a subdomain such as ai.yourdomain.com to the VPS, publish Open WebUI through a reverse proxy with an HTTPS certificate, and keep the Ollama API port closed to the public. Locations are available in Europe and America, so pick the one closest to your users. Set up a reverse proxy

• Keep the Ollama API private
• Publish Open WebUI over HTTPS
• Choose a location near your users

Ollama Cloud and managed AI services run models on someone else's infrastructure and bill by plan or usage. On your own VPS, Ollama and Open WebUI run on resources you control: you pick any open model, prompts and files stay on your server, and root access lets you add your own tools. Both apps are open source and free to use, so you pay for the VPS. In return, you handle updates yourself.

• Free choice of open models
• Prompts and files stay on your VPS
• You manage updates yourself

Pull the latest images and redeploy the stack in Coolify to update both apps. Chats, users, settings, and downloaded models live in persistent Docker volumes, so they survive the redeploy. With ZapX Pass, you can also create manual Backups of the VPS in the panel and restore them. Managing Backups

• Redeploy in Coolify for new images
• Data kept in persistent volumes
• VPS Backups and restore with ZapX Pass