how to configure Ollama in OpenWebUI

[harryz]

Hi guys,
Could you please guide me how to configure Ollama in OpenWebUI so that I can download a model to test?
I am having below error after I installed OpenWebUI.
I tried to change localhost to my OpenWebUi domain but still same error. My OpenWebUI domain is behind cloudflare proxy.

[james]

Hello @harryz
The OpenWebUI Cloudron app no longer comes bundles with an internal ollama server.
Sorry, the documentation is outdated and I will update it accordingly.

---

Please install the Ollama app.
Inside the Ollama app web terminal you can run this command to pull a modell:

ollama pull tinyllama

In the Ollama app web terminal also run:

cat /app/data/.api_key

Note down this key, you will need it.

---

Then you need to configure the OpenWebUI app.
Login as an admin and navigate to https://YOUR.DOMAIN.TLD/admin/settings/connections:
Press the configuration button for Ollama:

Click on local to change it to external:

Now when it shows external (1) configure the URL to your Ollama app url and paste the key from before into Auth

You can press the 'reload' icon to verify the connection:

If you see the success message in the top right, you should be good to go and click the Save button.

Now you can use your own Ollama server and modells in OpenWebUI:

[harryz]

Thank you, James. You saved my day. However, how can I increase system memory for OpenWebUI. I already resize Memory Limit for OpenWebUI app to 16GB but system memory in this case is a different setting?

[harryz]

My Ollama app has 12GB Memory.

said in how to configure Ollama in OpenWebUI:

Thank you, James. You saved my day. However, how can I increase system memory for OpenWebUI. I already resize Memory Limit for OpenWebUI app to 16GB but system memory in this case is a different setting?

[IniBudi]

@harryz is it fast using 16GB for generating answer in Ollama?

[james]

Hello @harryz
How much RAM does your server have?
How much RAM is used already?

In the ollama web temrinal run:

free -h

and post the output please.

[harryz]

Hi James, please refer below, thanks.

@james said in how to configure Ollama in OpenWebUI:

Hello @harryz
How much RAM does your server have?
How much RAM is used already?

In the ollama web temrinal run:

free -h

and post the output please.

[james]

Hello @harryz
There is your problem.
The model requires 8.7 GB and your system only has 7.9 GB of free memory left.