AI Integration local Ollama

Docs
,
1

I wanted to configure my local Ollama / Open-Webui AI Bot, but it wont connect to check for resources / LLMs - Anyone got that to work. I tried with the internal docker IP and the external one HTTP and HTTPS - All works fine frm that host using wget… Any tips?

2

Hello @walterh ,

The issue is that Ollama by default listens only on 127.0.0.1, so Docker containers can’t reach it.

  1. Configure Ollama to listen on all interfaces:
    sudo mkdir -p /etc/systemd/system/ollama.service.d
    sudo nano /etc/systemd/system/ollama.service.d/override.conf

  2. then add these:
    [Service]
    Environment=“OLLAMA_HOST=0.0.0.0”

  3. Reload systemd and restart Ollama:
    sudo systemctl daemon-reload
    sudo systemctl restart ollama

  4. Verify the setting is applied:
    systemctl cat ollama.service

In Document Server, go to AI tab → Add AI Model:
URL: http://172.17.0.1:11434/v1

172.17.0.1 is the host IP from Docker’s perspective. Don’t use localhost - it won’t work from inside the container.

Let us know if this helps.

3

Many Tahnks Nurullo!

That helped :slight_smile: as I was looking to my setup a bit deeper. It wasnt really that my Ollama daemon was not listening to 0.0.0.0, but I forgot that I was running ngnix reverse proxy.

Pointing OnylOffice to the external IP using https did the trick.

Thanks and regards!

4

Glad it worked! Thanks for sharing!