Skip to main content

​
Overview

Open the image model list in the creation panel on a video detail page and pick Meta Muse Image (marked new). It follows multi-step prompts and can keep editing an existing image, which is useful for covers, posters, and titles that have to stay readable.

​
Pain Point

I need a graphic with a clear title, and I want to tweak details from a reference instead of redrawing the whole thing every time. Switching models used to break the look.

​
Key Features

  • Select Meta Muse Image directly from the AI image model dropdown
  • Follows complex, multi-element prompts instead of flattening them
  • Edit a generated or uploaded image in place—no need to start over
  • More stable text rendering for titles and short lines on posters
  • Works with aspect ratio and 1K / 2K resolution settings

​
Highlights

Stay in the same image panel: generate first, then refine with a reference image instead of jumping to another tool.

​
Use Cases

  • Video-summary covers where the title must be readable
  • You already have a draft or screenshot and only want to change the background, props, or look
  • Long prompts with many elements that need to be followed in order

​
Tips

  • Write the exact copy and subject in the prompt; “make it nicer” is not enough
  • To edit an existing image, attach the reference first, then choose Meta Muse Image
  • Set ratio and sharpness in advanced settings before you generate

GPT Image 2 Generation

AI Image Advanced Settings

Xiaohongshu Image Generation

Try it now: Meta Muse Image Upgrade to Pro: More image runs and sizes → View Pricing