Run large language models with a simple API. Nuvlo installs it on a server in your own cloud account, with a domain, HTTPS and backups.
- Category
- AI
- Replaces
- OpenAI API, Anthropic API
- Server size
- HeavyNeeds real memory. Plan for a mid-size server or one of its own.
- Good for
- Private AI features without sending data out
- Open models
- Llama, Mistral, Gemma, Qwen and many more, pulled with one command.
- Simple HTTP API
- Generate and chat endpoints your own apps can call.
- Pairs with Open WebUI
- Install Open WebUI next to it for a chat interface.
Each app is its own Docker Compose project behind Nginx, with generated passwords and a Let’s Encrypt certificate.
- OllamaModel runtime and API
Ollama is an API with no web interface. Pull a model on the server after install, then pair it with Open WebUI for chat.
How do I install Ollama?
Pick one of your servers, choose Ollama from the one-click catalog and give it a domain. Nuvlo starts the containers with generated passwords, points the domain and issues the certificate.
What size server does Ollama need?
Heavy. Needs real memory. Plan for a mid-size server or one of its own.
What happens after install?
Ollama is an API with no web interface. Pull a model on the server after install, then pair it with Open WebUI for chat.
Is Ollama backed up?
Yes. Apps are sites like any other, so they use the same backup schedules and your own storage bucket.
Who updates Ollama?
You do. Ollama keeps the version it was installed with; to move to a newer release, pull the new image on the server over SSH. The dashboard has no update button for apps.
Does Nuvlo charge per app?
No. You pay your Nuvlo plan and your cloud provider for the server, and Ollama itself is free to self-host.