Discover the Best AI Image Generator: Create Stunning Ghibli Style Art with Liblib AI

Recently, AI-generated images reminiscent of Hayao Miyazaki's comic style have taken social media by storm! Whether on Xiaohongshu, Douyin, Weibo, or moments on WeChat, you can see scenes like floating mechanical castles, glowing forests, girls riding wooden bicycles, etc., all accompanied by the familiar soft tones of Studio Ghibli, instantly transporting you to a magical fairy-tale world.

Studio Ghibli's animation style is renowned for its soft color palette, vivid character expressions, and imaginative scene designs. This style is not just a visual language but also an emotional expression. In Hayao Miyazaki's works, nature and fantasy intertwine, with gentle and delicate lighting, creating characters full of spirit. These qualities have made the Ghibli style a unique cultural symbol, embodying countless people's longing for beauty and warmth. With Liblib AI and DeepSeek, creating Ghibli-style works is quite simple; I just need to ask DeepSeek to provide me with prompt words for generating Ghibli-style Miyazaki comic images, which is quick and convenient. For example, I typed into DeepSeek's question box: "Please help me provide prompt words for generating Ghibli-style Miyazaki comic images." DeepSeek directly provided me with several example prompts.

I simply need to copy these prompts into Liblib AI. For instance, here is a prompt based on the theme of Castle in the Sky: "A massive floating island covered with ancient ruins and dense vegetation, surrounded by steam-powered airships. A young girl in a blue dress stands on a stone balcony, her hair flowing with the wind, next to a small robot holding a flower. The golden sunset casts long shadows on the moss-covered stone pillars." Ghibli Studio style, Hayao Miyazaki's color palette, dreamlike atmosphere, highly detailed background, 8k resolution.

For the Liblib AI large model (checkpoint), I choose the Flux base F.1, adding LORA "Ghibli Style Comprehensive/Miyazaki Comic Generalization, weight 0.8." The prompt also needs to include the trigger phrase: "ghibli style da quan," followed by the prompts provided by DeepSeek, as shown above. The other parameters do not need to be changed; clicking "Start Image Generation," let's check the results together.

Using the same method, if I want a Totoro forest theme, I input the prompt: "ghibli style da quan + prompt (Under the moonlight, fireflies dance in the forest while two children dressed in raincoats wait under a giant camphor tree. A huge furry creature (Totoro) holds a leaf umbrella, its belly rising and falling as it sleeps. In the background, a traditional Japanese farmhouse is visible, with warm yellow light in the windows. Ghibli Studio style, magical realism, soft focus on foreground characters, nostalgic summer night atmosphere)," and importantly, don’t forget to include the trigger phrase.

Let's look at the results directly. If you want to change the aspect ratio (resolution), that’s also possible; just adjust the width and height. You can set it to a 9:16 or 16:9 aspect ratio, but remember, the Flux model has a minimum requirement of 1024×1024. I won't go into detail here, but feel free to give it a try.

Creating Ghibli-style Miyazaki comics is this simple. Have you learned how to do it? Hurry up and open Liblib AI to give it a try. No need to worry about prompts; let DeepSeek help you. It’s easy, quick, and practical! If this has helped you, please give it a like, and don’t forget to follow for more. You never know when you might need it! Stay tuned for the next exciting issue!