Use a clear, front-facing photo in good light with no sunglasses and no heavy filter. For figurine, Ghibli and DP prompts, a plain background gives the cleanest result. If the face changes too much, add the line “keep the exact same face and hairstyle” at the start of the prompt and generate again — one retry usually fixes it.
Related Articles
Photo first, video second
Every video prompt on this page works best in two steps. First create a strong still photo with the matching photo prompt, then upload that photo into Kling, Runway or Luma with the video prompt. Direct text-to-video skips the face-consistency that makes baby reaction and transformation videos feel real and shareable.
Read article →Post it so the algorithm picks it up
Export in 9:16, keep the first two seconds as the strongest frame, and use the trending audio that is already rising — not already peaked. Write the prompt name in your caption, for example “3D Figurine Toy Box prompt”, so search viewers can find you. Reply to the first ten comments quickly; early replies are the cheapest reach you will ever get.
Read article →Why baby reaction prompts go viral
Baby reaction videos work because the emotion changes inside ten seconds — surprise, then laughter. Keep the same baby face in every clip, shoot the prompt in a bright room setting, and never change the outfit mid-prompt. Consistency is what makes viewers follow the series instead of just liking one video.
Read article →