Lesson Content

Image prompting basics.

In Midjourney, adding an image URL at the beginning of a prompt uses that image as a visual reference that influences the generation. The image URL is added before the text prompt: [image URL] [text description] [parameters]. The text prompt modifies and extends the image reference rather than competing with it. Image prompting guides compositional and stylistic elements from the reference image.

The –iw (image weight) parameter.

--iw controls how strongly the image reference influences the generation relative to the text prompt. --iw 0.5 (low) – text prompt has more influence; the image is a loose reference. --iw 2 (high) – image reference has strong influence; results stay closer to the reference image's visual character. Default is approximately --iw 1. Adjust based on how closely you want generation to follow the visual reference.

What image prompting guides.

Image prompting is most effective for guiding: overall compositional structure, color palette and mood, general visual style and aesthetic character, and subject orientation. It is less precise for: exact subject appearance (see Omni Reference, –oref, for character and object consistency), specific fine details, or producing near-identical outputs from the reference.

Selecting reference images.

Effective reference images: clearly represent the intended visual direction, have a consistent style that translates to other subjects, and are high quality and high resolution. Poor reference images: visually ambiguous, contain multiple conflicting styles, or are low resolution with insufficient visual information for the model to reference.