Prompt Enhancer for Z-Image Turbo, using a local Ollama instance.
Simple command-line utility that will utilize a local Ollama instance to enhance prompts for Z-Image Turbo, using the official suggested enhancement template. Prompts must be supplied in a text file (one prompt per line).
I get good results on a 12GB RTX 3060 with a Q4_K_M quantized Qwen 3 2507 model; anywhere from a couple seconds to 1 minute to process each prompt, depending on length.
Make sure Ollama is running, then:
python enhance.py --prompt_file <filename containing your prompts, one per line> --model <ollama model to use>
For additional options, run python enhance.py --help.