How to Make a Cinematic Kissing Scene with Two AI Girls
Step-by-step guide to generating emotional, film-like intimate scenes with AI using text-to-image and image-to-video
Step 1 — Create the Starting Image
If you don't yet have an image where both characters are together in the same location, you can generate it easily.
Open Text-to-Image: https://app.zencreator.pro/tools/generator-by-prompt
For the example video, the chosen model was Nana Banana.
Here is the sample prompt for the starting frame:
A cinematic shot of two young adults standing closely inside a narrow, warm-toned elevator. Soft amber lighting from above creates gentle shadows on their faces. The girl touches the girl's face with a tender, emotional gesture. They stand very close, facing each other, with an intimate, dramatic atmosphere. Minimalistic red-brown elevator walls, shallow depth of field, film-look color grading, moody lighting, natural poses, expressive eyes, realistic skin texture, high-resolution portrait.
Choose:
- Your aspect ratio (9:16 for Reels / 16:9 for YouTube / 4:5 for Instagram)
- The number of image variations
Click Generate.
You can modify the prompt to change:
- Lighting
- Mood
- Characters
- Environment
- Emotional tone
Step 2 — Animate the Scene Using Image-to-Video
Open Image-to-Video: https://app.zencreator.pro/tools/video-generator
Select:
- Model: Seedance Pro Fast (optimal for natural, emotional, human-like motion)
Upload the image generated in Step 1.
Now describe the movement you want your characters to make.
Here is the exact sample prompt used in the reference video:
Cinematic video of two young adults kissing very passionately and touching each other extremely close inside a narrow warm-toned elevator. The atmosphere is filled with emotional tension. Their faces are inches apart, eyes locked, breathing softly, creating a powerful intimate moment. The girl gently touches the boy's face. Warm top light casts dramatic shadows on red-brown walls. The camera performs a slow dolly-in, then rotates around them with smooth handheld motion, adding intensity. Shallow depth of field, soft bokeh, warm amber color grading, film grain, natural skin texture, high-end cinematic look.
Click Generate.
Your video is ready.
Customize and Improve Your Result
You can:
- Extend the video by screenshotting the last frame and generating the next segment
- Write more detailed or more subtle action prompts
- Try alternate models for different cinematic styles
- Change camera movement (dolly in, rotation, slow zoom, handheld, static)
- Adjust light, mood, and intensity
- Experiment with different emotional tones (soft, dramatic, romantic, stylized)
Tips for Best Cinematic Quality
Use warm, soft lighting
Elevators, hallways, and evening rooms work particularly well.
Describe facial expressions
The model responds strongly to words like: "breathing softly," "emotional tension," "gentle," "natural skin texture."
Add camera motion
Even simple movement (slow dolly-in) makes the scene look premium.
Keep the environment simple
A minimal location (like an elevator) increases realism.
Maintain aspect ratio consistency
Match your image generation and video generation ratios.