What if a gigantic animal suddenly appeared in the middle of your favorite city? 👀 For today’s challenge, take any city and imagine it being turned into a surreal giant-animal scene. 🐱🌆 Try something like: 🐱 A giant cat using skyscrapers as a scratching post 🐘 An elephant walking between city blocks 🦅 A massive eagle perched on top of the skyline The key is to make it look as realistic as possible while keeping the city recognizable. Focus on scale, lighting, shadows, and realistic interaction between the animal and the environment. I used Nano Banana Pro for this one. Drop your original prompt + final image below 👇 Let’s see what wild combinations everyone comes up with. 🤨🔥

What if you could step into Minecraft, take any photo of yourself and turn the world around you into a Minecraft-style voxel environment. I used Nano Banana 2 for this one. Your task: 1. Upload a photo of yourself 2. Transform the environment into a Minecraft world 3. Replace [Your Name] 4. Share your result below 👇 Let’s see who can make the most convincing real person × Minecraft world.

I wanted to see how Kling 3.0 and Wan 3.0 compare when generating the exact same video. For this test, I kept everything the same: Same prompt, 720P, 8 seconds, Same reference image. The goal was simply to see which model handles the same generation task better. My quick take: Kling 3.0: → Stronger overall visual quality and motion consistency → More cinematic-looking results → Audio generation is a nice bonus → More expensive Wan 3.0: → Surprisingly competitive for the price → Good motion and image quality → Worth considering when generating lots of variations → Much more cost-efficient Which one would you pick? 👇

I swapped my character into this fight scene, and I think the result turned out way better than I expected. The movement feels surprisingly natural, especially considering I only used my character as the image reference and the fight scene as the motion reference. Genjutsu handled the motion and camera really well while keeping my character’s identity and outfit consistent. Definitely want to experiment more with this workflow 👀

The crazy part is how much you can do with this style once you get the right prompt structure. Different dishes, camera angles, little crochet details… it’s surprisingly fun to experiment with. Here’s full tutorial: https://youtu.be/MpyCkgEE6Uc?si=0SHR99cE6s-nfoDG Would you guys want to try it?