You Saw Something You Couldn't Explain. Good.
Most people assume cinematic AI video requires expensive tools, technical expertise, or a production budget. It doesn't. This tutorial walks through the exact workflow for creating photorealistic cinematic video using one free tool and a phone. No camera crew. No studio. No editing software. Just Higgsfield and nine steps.
Marlon Brand
Founder, Undeniable · Last updated March 2026
That feeling you just had watching that video...
That little pause where your brain went “Wait... what?... Wow.”
That's not an accident.
That's what happens when something breaks the pattern your eyes have been trained to expect.
And you just watched it happen with one free tool and a phone.
Here's exactly how.
The Full Breakdown
Record Your Reference Video
Record yourself doing whatever motion you want the final video to have. Walk down a hallway. Turn your head. Lift a coffee mug. Whatever.
This is your motion blueprint. Everything else gets built around it.
Your reference video. Phone quality is fine...the AI only needs your motion.
Grab a Reference Image
Screenshot a frame from that video. Doesn't need to be perfect. Just needs to show the positioning and environment clearly.
Open Higgsfield > Image Tab > Nano Banana Pro
This is where you build the scene you actually want to be in.
Imagine it fully before you touch anything. For us it was a detective standing in a narrow corridor... crime scene photos on the walls, red string connecting evidence, dim blue-grey lighting. Cinematic. Moody.
Higgsfield Image tab. Select Nano Banana Pro...it's the best 4K image model.
The generation interface. You'll upload your reference image here next.
Upload Your Reference Image
Hit the + button in the top left. Upload that screenshot.
Your reference image uploaded. The AI will use this for positioning and layout.
Prompt It
Use this structure...
“Positioning, body movement and room layout should be exactly like the reference image. [Describe your character in detail]. [Describe the scene]. [Lighting and atmosphere]. Cinematic. Ultra-realistic.”
Here's what we used:
“Positioning, body movement and room layout should be exactly like the reference image. A heavyset man with a short beard wearing a black cowboy hat stands in profile inside a narrow corridor, holding a coffee mug close to his face. He is wearing a dark navy slim-fit blazer over a white dress shirt and dark fitted trousers. The corridor walls are covered in crime scene photographs, newspaper clippings, yellow sticky notes, and red string connecting evidence in a chaotic web. Dim, moody blue-grey overhead lighting. Deep shadows. Intense brooding focus. Cinematic crime thriller atmosphere. Ultra-realistic.”
Hit generate.
Regenerate If It's Off
If the image comes out distorted or wrong... regenerate. This is normal. Keep going until the positioning and feel are right.
Distorted. Regenerate.
That's the one.
Upscale It
Once you have the image you want...
- →Click “Open In” at the bottom right.
- →Select “Upscale.”
- →Click x2 and hit Upscale.
Download it. Now you have two files... your original reference video and your upscaled AI image.
Click “Open In”
Select Upscale
x2 upscale. Download when done.
Open Kling 3.0 Motion Control
Go to the video tab. Select Kling 3.0 Motion Control.
- →“Add Motion to Copy” tab... upload your original reference video.
- →“Add Your Character” tab... upload your upscaled image.
Very important: Enable Scene Control and set it to Image.
Video tab. Kling 3.0 Motion Control...this is where motion meets image.
Upload slots ready
Both uploaded. Scene Control: Image.
Prompt the Motion
“Camera stay still. Copy exactly the motion from reference. No unnecessary movements added. Realistic movements only. No morphing, no AI distortion. Photorealistic human motion only.”
Hit generate.
Done.
Your motion. Your scene. Your character. Something that looks like it cost thousands... made with Higgsfield and a phone.
The final output. 27 seconds of cinematic video. One tool. One phone.
Now Think About What You Just Did
You just created a cinematic ad creative... without a camera crew, without a studio, without a budget.
Six months ago this wasn't possible.
Six months from now everyone will know how to do it.
Right now... almost nobody does.
That window is open. And it's closing.
We're building something for the people who want to stay ahead of it.
Already Running Meta Ads?
We built an agent for that. It's live right now.
It writes your Meta ads across all five awareness levels. Writes your landing page copy. Writes your emails. Gives you a dual analysis report and decision-making framework for every campaign.
One system. Installs in 60 minutes. Built from $1.6M+ per month in Meta ad spend.
The Creatives Agent
What you just learned in this tutorial... that's the manual version.
We're building an agent that creates ad creatives on command. Pulls from research. Iterates off your winning angles. Tests new concepts based on what's already working.
It's not live yet. But it's coming soon.
Join the Creatives Agent Waitlist
The people already inside will get first access when it drops.
Frequently Asked Questions
More from the Lab
Clean Copy...Make AI Text Sound Human
Free browser tool that strips AI tells from your copy. Paste, pick a mode, get human output.
Try it →--meta-adsMeta Ads Credit Card Deadline: What to Do Before April 1st
How to switch billing, what it means for cash flow, and the exact message to send each type of client.
Read →--ai-toolsControl Your Computer From Your Phone Using Claude
Step-by-step guide to Claude Code Remote Control. No coding required. Set up in 10 minutes.
Read →--ai-toolsClaude vs ChatGPT for Coaches and Consultants
Which AI actually does the work? Architecture, persistence, and compounding compared.
Read →Want to See What AI-Powered Marketing Looks Like?
We use AI to build, launch, and scale faster than any traditional agency. One new client per quarter.
// the window is open... it's closing
Apply to Work With UndeniableNo commitment. No pitch deck. Just a straight conversation about what's possible.