Get image prompt from OpenAI or a compatible API. By default gpt-3.5-turbo is used.
gpt-3.5-turbo
Compare the results of DALL·E and local SD models to see which one is better at generating images from text.