COMMUNITY PAGE
Cinematic E-Commerce Backpack Ad (Step by Step)
Workflows Used
floyoofficial
1.8k
gemini 3 pro
Image
Image2Image
Image Editing
Nano Banana Pro Edit
Edit any image with text using Nano Banana Pro, Google's Gemini 3 Pro image model. Upload your image, describe the change, add references, and hit run.
Nano Banana Pro · Edit Any Image
Edit any image with text using Nano Banana Pro, Google's Gemini 3 Pro image model. Upload your image, describe the change, add references, and hit run.
Kling 2.5 Image to Video
Step 1: Collect Reference Images
Before writing any prompt, I first collect reference images.
References act like a visual guide for the AI.
I usually separate them into:
Face reference – to lock the model’s identity
Backpack reference – exact bag, color, material
Environment reference – forest, jungle, city, etc.
Pose references – walking, sitting, holding a camera, resting
Each reference has only one job.
Mixing everything into one image usually creates confusion later.
Step 2: Create the Character + Backpack Base
I use Nano Banana Pro to lock the basics.
Prompt for Model: Realistic character reference sheet, young slim male with curly black hair, forest-green retro polo with cream stripes, khaki trousers, white-green sneakers. neutral relaxed pose.
layout: full-body turnarounds (front, 45°, side, back) + head angle grid (front, 3/4, profiles, up, down) with subtle expression variation.
style: soft studio lighting, warm beige background, cinematic realism, subtle film grain, sharp facial and fabric detail, accurate anatomy.
Prompt for backpack: backpack reference sheet, burnt-orange canvas + dark brown leather straps + brass buckles.
layout: row1 full turnarounds (front, 3/4 left, left, back, right, 3/4 right)
row2 worn on simple mannequin (front, side, back)
row3 close-ups (logo patch, buckle, zipper, canvas texture, stitching)
style: clean studio beige background, soft shadows, consistent lighting, cinematic realism, slight film grain
quality: sharp details, accurate colors, no text, no hands, no props
Once both are locked, I can place the same character and the same bag into any scene without things changing randomly.
🔗 Nano Banana Pro workflow:
https://www.floyo.ai/workflows/nano-banana-pro-edit-image-to-image-akmj3e2cdrb1
Step 3: Brainstorm the Story (You Can Use ChatGPT Here)
Even short videos need a simple story.
You don’t need to overthink the story.
Before turning images into video, I quickly plan the flow of the short.
This is where ChatGPT is very helpful.
I use it to:
Brainstorm story ideas
Decide shot order
Plan transitions
Think of small moments (walking, stopping, resting, shooting photos)
Even a simple story works:
walk → stop → rest → open bag → take photo → continue journey
Having a rough story makes the visuals feel connected.
Step 4: Generate Scene Images
Now I start generating scene images using Nano Banana Pro.
For every image:
I use face reference
I use backpack reference
I use environment reference
Pose reference (only if needed)
Prompt Sitting Pose Scene:
Use image 1 for face reference.
Use image 2 for backpack model and color.
Use image 3 for the forest environment.
Use image 4 for sitting pose reference.
A young man sitting on a fallen tree in the forest. Backpack placed beside him. Natural daylight, cinematic depth of field.
I keep prompts short and clear. That works best.
Generate Multiple Story Frames
I repeat this step to create different moments:
Walking through the forest
Sitting and resting
Holding a camera
Standing and observing
Each image feels like a frame from a film.
🔗 Nano Banana Pro workflow:
https://www.floyo.ai/workflows/nano-banana-pro-edit-image-to-image-akmj3e2cdrb1
Step 5: Convert Image to Video
Once I have all the images, I convert them into video using Kling 2.5 (Image-to-Video).
I keep video prompts very simple to avoid weird motion.
_1767907335389.gif)
Prompt: a young man with curly hair wearing a green short-sleeve shirt, beige pants and an orange backpack walks forward through a sunlit green forest. camera stays in front of him at chest level, moving backward smoothly to keep his face in frame as he walks. natural forward footsteps, relaxed arms, steady breathing, cinematic sunlight through leaves, warm colors, sharp backpack logo, realistic motion, no backward walking.
_1767907205457.gif)
Prompt: A young curly-haired man in a green polo and beige pants walks slowly along a mossy forest trail, seen from behind. an orange canvas backpack sways naturally on his back as his hand brushes a damp moss-covered rock. soft morning mist moves through tall trees, sunlight filtering between branches. the camera follows at waist height with gentle handheld motion, leaves passing in the foreground. calm cinematic travel mood with natural, realistic movement.
Sometimes, when I need a very specific shot, I have to use Kling 2.5 start and end frame workflow.
_1767907927613.gif)
Prompt: a young curly-haired man wearing a green polo and beige pants walks beside a rocky forest river during golden hour, warm sunlight and tall pine trees. he stops near the river and gently slides his orange canvas backpack off his shoulder with both hands. he lowers the backpack onto a mossy stone, his hands still holding the straps. he pauses, adjusts the position of the bag with small realistic movements, then slowly unbuckles the leather straps one by one, lifts the top flap in a natural motion, revealing the DSLR camera inside. he leans forward slightly and reaches into the open compartment, fingertips touching the camera grip, beginning to pull it out. realistic pacing, micro hand movements, no jump cuts, smooth natural transitions, cinematic handheld camera, warm soft light, consistent look and color.
🔗 Kling 2.5 Image to Video workflow:
https://www.floyo.ai/workflows/kling-2-5-image-to-video-eashmyqv54ck
Step 6: Final Edit in After Effects
In After Effects, I:
Arrange all clips
Add smooth transitions
Do basic color grading (If needed)
Add sound
This step turns separate AI clips into one cohesive cinematic piece.
Final Thoughts
You don’t need perfect prompts or complex setups.
You just need clear references, patience, and a clean process.