Unable to configure ollama on fresh D11 site

Problem/Motivation

I'm using DDEV and am unable to configure the provider to use a model I've installed in the ollama container.

When I try to configure it via /admin/config/ai/providers/ollama, I get the following error:

Failed to get models from Ollama: Client error: `GET http://host.docker.internal:11434/api/tags` resulted in a `404 Not Found` response: 404 page not found

This is regardless of if I set the Host name to "http://ollama" or "http://host.docker.internal" (both with port 11434.)

From the command line, if I run "ddev exec curl host.docker.internal:11434", I get "404 page not found".

But, oddly, I can utilize the API Explorer successfully against the installed model and it works fine.

Steps to reproduce

• Fresh Drupal 11 site.
• Dev versions of both AI and AI Provider Ollama modules.
• Using the docker-compose.ollama.yaml file from the AI docs folder.
• Install the gemma3:1b model via "ollama run gemma3:1b"
• Configure the provider with either "http://ollama" or "http://host.docker.internal" with port 11434.
• I can successfully use /admin/config/ai/explorers/chat_generator to get a valid response.
• When I try to use the "AI Content Suggestions" module to use "AI Suggestions" form widget settings (to suggest a title), I get an "There was an error obtaining a response from the LLM." in the UI and the following in Recent log messages, "Error invoking model response: POST predict: Post "http://127.0.0.1:43741/completion": EOF" (I have no idea where 43741 is coming from.)

-mike