Ollama

About Ollama

Run Llama 3, Mistral, Gemma, Phi and other large language models locally and self-hosted with OpenAI-compatible API endpoints.

Deploying it

1. 1Open the Hub in your panel and pick Ollama. Its detail view shows the image, port, volume and environment it ships with.
2. 2Click Deploy. The wizard arrives with Source and Runtime filled in; review the environment and set CPU and memory limits if you want them.
3. 3Review and deploy. Traefik routes it on an automatic URL under your wildcard domain, and it becomes an ordinary service with its own logs, backups and rollbacks.