Midjourney Creators Share Style Reference Codes and Workflows

From July 11 to 13, multiple Midjourney creators shared reusable style reference codes and workflows for the `--sref` parameter on X. This feature allows users to consistently reproduce specific visual styles by locking onto a numerical code, eliminating the need to repeatedly guess manual prompts, making it a popular image generation trick recently.

Key Parameters and Codes

Creators shared various `--sref` codes for different styles: @michaelrabone provided `--sref 2916766707` (samurai/oni mask/hologram characters) and `--sref 3889935155` (geometric low-poly, submarine hybrids, cross-section infographics, industrial mechanical designs, vibrant colors). @azed_ai shared `--sref 1974399503` and `--sref 1926423158`, while @umesh_ai used `--sref 3911489532` on Midjourney 8.2 Preview. Common parameter combinations include `--chaos` (values ranging from 8 to 25), `--stylize`, `--ar` (e.g., 3:2, 16:9), `--profile n1y63rs`, and `--sw`. @michaelrabone also demonstrated cross-subject style reuse by applying the same `--sref` to both Chihuahua and French Bulldog prompts.

Usage Tips

@OVolosin82152 pointed out that the styles in these images aren't guessed via manual prompts but are achieved using `--sref`. @michaelrabone added in multiple posts that `--stylize` can boost creativity, while `--chaos` increases result diversity, and combining them helps fine-tune the ideal output.

2026-07-11 ~ 2026-07-13 · 8 related posts