Ollama
A local Ollama server is included in the package and is enabled by default. It listens on localhost port 11434. To disable local Ollama and use an external one edit…
Open WebUI is a self-hosted web interface for local and remote LLM backends, with multi-user accounts and a ChatGPT-style experience.
DANIAN runs Open WebUI as a fully managed service: installation, updates, backups, monitoring and security are handled for you.
See the app catalog for plans, pricing and a free trial.
A local Ollama server is included in the package and is enabled by default. It listens on localhost port 11434.
To disable local Ollama and use an external one edit /app/data/env.sh using the File manager.
export LOCAL_OLLAMA_ENABLED=false
export OLLAMA_API_BASE_URL="https://<remote-ollama>/api"
Be sure to restart the app after making any changes.
Model Storage
By default, models are stored at /app/data/ollama-home/models and are part of the app backups. Models can be very large and it's probably unnecessary to backup models.
Backup of models can be skipped by changing the storage directory to a Block storage volume.
- Add a volume
- Add the volume to the app mounts
- Set the model storage in /app/data/env.sh using the File manager
export OLLAMA_MODELS=/media/ollama-vol/models
Be sure to restart the app after making any changes.
Model Library
Ollama supports a list of models available on ollama.com/library. Sample sizes and memory requirements of models is available here.
Find more information about our company, platform, and services on our website.
Visit danian.co
