1.Due to platform restrictions, it is not possible to use restricted prompt words or generate restricted images. 2. The generated image lacks clarity 3. It is difficult to reproduce the details required for the art of maru-swallowing, such as prey being violently squeezed by the stomach wall.

However, the performance of this model in terms of character consistency is truly astonishing. The steps I am about to present may have some shortcomings, and I look forward to your sharing of tips and suggestions for generating more expressive images and prompts.

I use the following URL to use nano-banana. https://lmarena.ai/

↑After completing the above steps, there is a chance that your image will be processed by nano-banana. If not, please re-enter the image and prompt word and try again.

Step 1: Find a character sprite. I’d like to manipulatiy character illustrations in Photoshop to achieve the look of the second image, featuring a flesh-colored background and a tilted figure. The flesh-colored background is because I think it can generate a better stomach wall image later. I think that tilting the characters preserves more of their individual traits. This is because the square images generated by nano-banana are all of a 1024*1024 size.

Step 2: Change the action of the character in nano-banana. Input the 3D drawing you want to process, and then enter the following prompt words: Keep the artistic style of the character unchanged. Keep the characteristics of woman's twintails, clothing and legwear unchanged. Let this woman sit down on the ground. Make the woman's two legs up and her two feet in the air. One of the lower legs is higher than the other. You can adjust your prompt based on the characteristics of the illustration or the areas where nano-banana is prone to making mistakes when processing your image. For example, in my second prompt, I specifically emphasized "Keep twintails unchanged." This is because the generated images always showed the character with their hair down (image 1). You can also modify the prompt words according to your preferred pose. I ultimately obtained image 1 and image 2, and finally combined them into image 3 in Photoshop. (Since I don't like the characters to take up too little space in the picture, I enlarged them.)

Step 2.5: Remove the shoes. (Those who like vore shoes can skip this step) The shoes of the character were removed through the following cue words: Keep the position and proportion of the woman in the picture unchanged. Make the picture with higher clarity. Remove the woman's boots. Make the color of her two soles matches the color of her pantyhose.

Two things worth noting: 1. When writing prompts, please try to use "woman" more often and "girl" less, otherwise it may be rejected by the platform due to triggering platform restrictions. 2. Please try to break down the steps for image processing, otherwise there may be a messy image. (For example, I separated changing the pose and removing shoes into two separate steps)

Step 3: Add mucus. Use the following prompt words: Keep the artistic style of the character unchanged. Keep the characteristics of woman unchanged. Make the woman covered with green mucus. Keep the position and proportion of the woman in the picture unchanged. Make the picture clearer. I like to overlay the generated image(image 1) with the original one to achieve a slimy and transparent effect (image 2).

Step 4: Change the background. Use the following prompt words: Change the background inside the narrow stomach environment, filled with green mucus. The surrounding area was also splashed with green slime.There are bumps on the stomach wall.There are also protrusions with holes on the surrounding flesh walls, spraying green mucus. Let this woman lay back agains the green mucus. It is possible to appropriately change a woman's posture to better integrate the character and background. You can adjust the details of the prompt words to make the generated image more in line with your preferences. With that, our picture is complete.

Welcome everyone to share your pictures and prompts. We also hope to see more excellent steps and prompts!