Midjourney Cheat Sheet

Sets image width-to-height ratio. Default is 1:1.

Tells Midjourney what not to include. More effective than saying "without" in the prompt.

Note: Equivalent to negative weight in multi-prompting (item1::-.5)

Determines rendering time and quality. Higher values consume more GPU time. Default is 1.

Sets the starting point for generation. Same seed + prompt = similar images.

Selects the Midjourney algorithm version. Latest is v6.

Controls result variability. Higher = more unexpected outcomes.

Applies alternative aesthetics. Raw offers less automatic beautification.

Controls strength of Midjourney's default aesthetic style.

Controls processing speed and GPU usage.

• fast: Default mode using GPU time allocation
• relax: Queue-based, no GPU usage (0-10 min wait)
• turbo: 4x faster but uses 2x GPU minutes

Amount of motion in the video. Default is low.

• Low: Still scenes, slow motion, subtle movements
• High: Big camera motions, larger character movements

More precise motion control for video prompts.

Selects the Midjourney style reference algorithm version. Required for most sref codes.

Sets the strength of style influence when using --sref. Higher = more similar to reference.

Uses an image URL as a character reference to maintain consistent appearance across scenes.

Sets characterization strength. Lower = face only. Higher = hair + clothing. Default is 100.

Adjusts influence of image prompts vs text prompts. Default is 1.

Generates images in an anime/manga style.

Makes the job private, not visible in public feeds.

Runs multiple jobs from a single prompt.

Ends generation early, resulting in less detailed images.

Generates images suitable for seamless patterns.

Saves a progress video of the initial image grid generation.

Introduces quirky and offbeat qualities to images.