I made this workflow because Z-Image Turbo works best with detailed prompts with 500+ words.This simple Workflow will enhance your Prompt using your favorite Ollama Model. Useful also if you use a multilingual Ollama model to translate from any language.<h3 id="version-2.0">Version 2.0</h3>So this is a major revamp of my previous versions.This now includes:<ul><li>Ollama Enhance Prompt</li><li>Use a Rereference image for either or both together of<ul><li>Ollama Reference Image (+ Prompt)</li><li>Controlnet</li></ul></li></ul><ul><li>Change a Person using a Character Lora, so you can have 2 different Characters in one image, what is otherwise not possible with ZIT (inspired by <a target="_blank" rel="ugc" href="https://civarchive.com/models/2232773/2-lora-characters-z-image">https://civarchive.com/models/2232773/2-lora-characters-z-image</a>)</li><li>SeedVR2 Upscaler</li></ul>There are notes in the workflow to explain most of the important things.Note:I use scheduler beta57, which is not by default available in "Detailer (SEGS)" note, but you can easily add it, simply add it to the "ADDITIONAL_SCHEDULERS" list in the file <code>custom_nodes/comfyui-impact-pack/core.py</code>, so the line read as:<pre><code>ADDITIONAL_SCHEDULERS = ['AYS SDXL', 'AYS SD1', 'AYS SVD', 'GITS[coeff=1.2]', 'LTXV[default]', 'OSS FLUX', 'OSS Wan', 'OSS Chroma', 'beta57']</code></pre><h3 id="version-1.1">Version 1.1</h3>Add option to use a reference image for Ollama (make sure to use a Vision Model).The first example image uses the second example image as reference, adding and changing details using Ollama.<h3 id="version-1.0">Version 1.0</h3>This is the base version.The first example image is made with Ollama Enhance, second is the same prompt without Ollama.

Z-Image Turbo + Ollama - Simple Workflow

Add option to use a reference image for Ollama (make sure to use a Vision Model)

Files with SHA256: c53ba5a5928ecf006b145bc6b0a3ca3de3561e22ce59455b36d1735f499ea3ba