How It Works
The process is simple. You take a photo or a screenshot of yourself, generate a new image that places you in a completely different environment, and then turn that image into a video with realistic motion.
Here are the steps:
- Take a screenshot of the first frame of your video (or use any clear photo of yourself)
- Generate the image using Nano Banana 2 with an image prompt
- Generate the video using an AI video model like Kling 3.0 Omni with a video prompt
For the video step, you need to use Kling Omni to make it work, or other similar alternative. Kling 3.0 is strong with consistent character motion.
You can access these AI tools in platforms like Freepik, ImagineArt, Higgsfield, etc.
Also read: How to Create AI Videos Using ImagineArt Workflows
Below are 3 complete prompt sets: mountain cliff, ocean raft, and jungle. Each includes an image prompt and a video prompt.
Scene 1: Mountain Cliff Edge
Image prompt:
Same person, identical face, hairstyle, and body proportions, sitting in the exact same pose and camera framing as original (do not change posture or angle). Replace the environment with a high mountain cliff edge.
The rock surface is carefully shaped to match his sitting posture so he looks naturally grounded on the edge. One side drops into a deep valley with clouds below. His hands are naturally resting on the rock surface for balance.
Outfit changed to wind-blown outdoor clothing (jacket, rugged pants, boots). Strong wind effect on clothes and hair. Dramatic sky with moving clouds. Cinematic lighting, ultra realistic, high depth, 4K detail. Ultra wide angle.
Video prompt:
Camera locked, same framing and sitting pose. The man is sitting on the edge of a high cliff, maintaining the exact same posture. Strong wind continuously blowing his clothes and hair in one clear direction.
Clouds moving fast below in the valley, creating a deep sense of height. Many birds (eagles or large birds) flying across the scene at different distances, some far in the background, one or two passing closer to the frame.
Small dust particles and tiny rocks slightly shifting near the edge due to wind. The man slightly looks left and right, reacting naturally to the environment while maintaining pose.
Cinematic lighting, high depth, ultra realistic motion, no face or body distortion.
Scene 2: Raft in the Ocean Storm
Image prompt:
Camera placed in front of the man at a natural eye-level angle, not too close. A young man sitting on a small wooden raft in the middle of a vast open ocean. He maintains a natural sitting posture while holding a wooden plank or part of the raft firmly in his hands, using it to balance and slightly steer himself.
The raft is not static. It continuously moves and drifts with the ocean waves, slightly changing position and angle over time. Water flows and splashes naturally around the raft edges.
The man makes small realistic movements, slight body balance adjustments, subtle head turns, reacting to the environment. His grip on the wood looks functional, not posed.
Strong wind blows consistently, affecting clothes and hair. Ocean waves rise and fall dynamically. Dark storm clouds move across the sky.
Occasional subtle lightning flashes in the distance (not too frequent), creating brief natural light changes. No extreme effects. Cinematic, ultra realistic water physics, natural motion, grounded interaction, no floating effect, no distortion of face or body, smooth continuous movement.
Video prompt:
Front camera, eye-level view. A young man sitting on a small wooden raft in the middle of the ocean.
He holds a wooden plank and slowly paddles the water in a natural way. The raft moves forward gently with the waves and slightly rotates. Water splashes lightly where the wood touches the surface.
Wind blowing clothes and hair in one direction. Ocean waves moving continuously but not extreme.
The man naturally looks left and right and sometimes looks forward. Subtle head and eye movement, no exaggerated motion.
Dark cloudy sky with very light thunder and occasional soft lightning in the distance.
Keep everything realistic and stable. No distortion, no extra motion, no disappearing objects. Smooth cinematic movement.
Scene 3: Jungle With a Black Panther
Image prompt:
Same person, identical face and body proportions, sitting in the exact same reading pose with book. Transform environment into a dense jungle with cinematic lighting. Outfit changed to rugged explorer style. A realistic black panther slowly walking behind him toward him. Maintain exact hand, book position and posture. Ultra realistic, depth of field, dramatic light rays.
Video prompt:
Camera locked, same framing. A young man sitting and reading a book in a jungle environment. He maintains the exact same sitting pose and position throughout.
He flips the book page exactly three times, each flip happening slowly and clearly at regular intervals (approximately every 3 seconds). Each page turn is natural, with visible hand movement and realistic page motion.
A black panther is present in the background from the beginning and slowly walks forward in a straight path without sudden changes.
Leaves move slightly with soft wind. Lighting remains stable and cinematic. The man stays calm and focused on the book, only minimal head movement.
No distortion, no flickering, no disappearing elements. Maintain full consistency of face, body, and objects. Smooth, realistic motion.
Tips for Better Results
The image prompt does most of the heavy lifting. If the generated image doesn’t look natural (wrong pose, weird lighting, body floating above the surface), the video will inherit those problems. Spend the time getting the image right before moving to video.
Read more: 7 New AI Tools That Will Save You Hours of Work
Pay attention to the lines about “no distortion” and “no face or body change” in the video prompts. These are there on purpose. AI video models tend to warp faces and shift body positions over time. Including these instructions reduces that.
The “camera locked” instruction is also important. Without it, the AI might add random camera movement that breaks the scene. Keeping the camera still makes the video feel more cinematic and keeps the focus on the environment.
You can swap the scenes for anything. A rooftop in a city, the surface of another planet, the middle of a battlefield. The structure stays the same: describe the person’s pose and position first, then describe the environment around them, then add the technical details at the end.
If you want to explore more AI tools, check our homepage. Or visit the Blog for more prompts and tutorials like this one.
See how I generate engaging YouTube thumbnails:
