Anima ReStyler — v1.2

Anima ReStyler — v1.2 creates a new image featuring the character from an input image. Provide a character image, target image size, and prompt; it can change the character’s style or place the character in a new environment.

For canvas extension, add a solid-color block to the input and mask only that area; the workflow handles image stitching and cropping.

The prompt `(split screen, multiple views:1.2)` directs attention to the left half of the image, while the Anima Edit LoRA provides Cosmos-Reference image conditioning for character consistency.

A character image in a neutral pose against a simple background is suggested as a starting input.

Suggestion · non vérifié

Use the `animaRestyler_v12.json` workflow file; the extracted file contains 82 nodes.

Model files listed for this workflow are: `anima_aestheticV11.safetensors`, `AnimeEditV2.safetensors`, `Extend Image (Anima Edit) v1.safetensors`, `qwen_3_06b_base.safetensors`, and `qwen_image_vae.safetensors`.

Node packs listed for this workflow are: `comfyui_essentials`, `comfyui-kjnodes`, `comfyui-prompt-control`, and `scale-image-to-total-pixels-advanced`.

Use prompt weights such as `(at night:2.0)` to emphasize a prompt element.

Use extended syntax such as `[:close eyes:0.3]`; the phrase activates after 30% of generation steps.

A character design description is optional and may help with trickier generations.

Suggestion · non vérifié

Different seeds can produce very different results; some can ruin a generation while others work well.

Monochrome images and sketches can be difficult, so colorize them first; changing the pose while keeping the style 100% consistent is also difficult.

Suggestion · non vérifié