Everything on Midjourney v6 and Niji 6 in one place: parameter deep dives, style control, character consistency, examples and guides.
Keep the same character across multiple AI-generated images. Use identity anchors, appearance anchors, wardrobe lock-in, signature props, LoRA training and Midjourney --cref/--sref. Includes wrong/right examples, five samples and weight tables.
Read full articleTeardown of a stylized illustration prompt — narrow palette, ink wash texture, traditional composition — with three rewrites (mural / fine line / modern minimal) and a Niji/SDXL parameter table.
Read full articleMidjourney --stylize controls how literally the model follows your prompt. 0 is the most literal, 1000 the most creative. This article gives recommended values per scenario with samples.
Read full articleMidjourney --chaos controls how different the four candidate images are. 0 is near-identical, 100 is wildly divergent. Practical workflow tips inside.
Read full articleMidjourney --weird (0–3000) pushes the model toward unusual interpretations. This article explains four reference values and how --weird differs from --chaos and --stylize.
Read full articleMidjourney --ar controls aspect ratio. This article covers 6 common ratios — square, e-commerce, portrait, illustration, banner, vertical — and how the ratio shapes composition.
Read full articleMidjourney --seed is the key to reproducing images and building consistent series. This article explains how to capture a seed, reuse it across prompts, and combine it with --cref.
Read full articleMidjourney --cref fixes a character with a reference image. This article covers --cw weight, combination with seed and --sref, and the full mini-series workflow.
Read full articleMidjourney --sref uses an image to lock the visual style. This article explains --sw weight ranges, multi-image stacking, and combination with --cref.
Read full articleNiji 6 is Midjourney's anime model. This article explains the four --style sub-modes, recommended --s values, and how to write prompts for Japanese, Korean and Chinese animation looks.
Read full article