Co‑Creating a Mother’s Day Animation with Midjourney AI: A Step‑by‑Step Guide

On Mother’s Day, NetEase Entertainment released a warm‑hearted animation titled “Now, Go Home”, which was co‑created with AI using Midjourney.

01 AI Script Description

Compared with traditional animation scripts, AI‑generated visuals require more detailed textual descriptions. The original script is expanded with specific keywords to form an AI description script.

The prompt framework includes:

Subject description : scene framing (camera angle, lighting, composition), e.g., “night, raining”.

Main subject : e.g., “a young girl”.

Detailed actions and attire : e.g., “wearing an orange‑red dress, walking alone at an empty traffic‑light intersection”.

Style description is derived from Midjourney’s

/describe

command, which generates textual prompts from reference images. Optimized style keywords might be “dreamy illustration, gray tone, whimsical children’s book style, misty landscape, watercolor, bold brushstrokes”.

Combining subject and style keywords yields the final AI script, for example:

“Night, raining, a young girl in an orange‑red dress walking alone at an empty traffic‑light intersection, dreamy illustration, gray tone, whimsical children’s book style, misty landscape, watercolor, bold brushstrokes.”

02 AI Drawing

Midjourney currently outputs PNG files only, so characters and scenes are generated separately to facilitate later compositing.

Character generation : Include terms like “white background, character design, multiple poses and expressions, animation‑style expression” to obtain layered character assets that are easy to cut out.

Example prompts:

Girl: “20‑year‑old female, orange‑red dress, white background, character design, multiple poses, whimsical children’s book style, watercolor, bold brushstrokes.”

Mother: “30‑year‑old Chinese woman, gray clothing, white background, character design, multiple poses, whimsical children’s book style, watercolor, bold brushstrokes.”

Scene generation : Upload up to 30 reference images (PNG/JPG) for Midjourney to analyse and incorporate visual elements. Example scene prompt mirrors the AI script above.

03 Post‑Processing, Color Grading, Layering

AI‑generated images often contain logical bugs; illustrators must manually correct these and separate foreground elements into layers for animation.

Final refined prompt for a specific frame might be: “A watercolor close‑up of a 5‑year‑old girl in an orange‑red dress sitting on a white paper airplane, blue sky, pink and gray tones, dreamy illustration, whimsical children’s book style, misty landscape, watercolor, bold brushstrokes.”

The practical benefits observed include:

Increased efficiency in early creative stages by quickly generating diverse concepts and styles.

Reduced communication overhead with stakeholders through visual proposals.

Significant time savings in mid‑production as AI creates many characters and backgrounds.

Expanded stylistic variety beyond the illustrator’s usual palette.

Limitations remain, such as occasional logical inconsistencies and difficulty rendering extreme angles, which still require manual touch‑ups.

Overall, Midjourney dramatically speeds up workflow, allowing more focus on creative direction while alleviating the initial anxiety around AIGC adoption.