Using ChatDev with local Ollama

[DN2048]

I'm trying to use ChatDev with a locally implemented Ollama service but ChatDev seems to just ignore the setting.

I have BASE_URL=http://localhost:11434 in .env file but in the frontend page, all the workflow examples either use Gemini or OpenAI and in the Model Provider I can't select anything else besides those two.

Is there a extra setting somewhere? Does it need a specific workflow?

PS: I know the Ollama service is up and running because I'm using it on other local services.

[zxrys]

#505
Please refer to this PR: Use http://localhost:11434/v1/ for base_url, select OpenAI as the provider, and then enter a random api_key.