Playbooks PromeAI

Train a Custom Consistency Model from One Image

Create your own consistency model using a single reference image and instantly generate high‑quality images for multiple concepts in under a minute.

Takes

Image Text

Produces

Image

What you'll produce

Input

Upload a colorful surreal painting of a floating island, set the prompt to "vibrant surreal", and add usage scenarios "sneaker", "fox", "modern car".

Output

Three distinct images: a futuristic sneaker with glowing accents, a stylized fox with neon fur, and a sleek modern car cruising above clouds.

The Workflow

7 steps · click a step number to mark it done

Open PromeAI and have it ready before you start

Click the Consistency Model option in the left‑hand sidebar

Press the image‑plus icon in the settings panel and upload the reference image you want to train on

Enter a brief style description in the Prompt field, such as "vibrant surreal"

Add usage scenario prompts for the concepts you want (e.g., "sneaker", "fox", "modern car")

Click the Train button, confirm the 4‑coin cost, and wait roughly 60 seconds while the model builds

Select the newly trained model, type a single‑word prompt for each scenario, and click Generate to receive a high‑quality image

Open PromeAI and follow the steps above.