0:00 / 0:00

How to make AI videos in REAL TIME 🤯🔥 @krea_ai gave me early access to this tool to test it out, pretty insane stuff! I think it’s cool to test around and play around with however I do feel like there is a lack of real world use case for this tool and its infancy stages. I look forward to seeing how it progresses. #KreaAI #RealtimeAIVideo #AIVideoGenerator #GenerativeAI #CreativeAI #AIForCreators #AIMagic #AIContentCreation #NextGenVideo #SmartAI #DigitalCreators #InnovativeAI #AIStorytelling #FutureOfContent #AIEditingTools #CreatorEconomy #AIVisuals #LiveAIVideo #FutureOfFilmmaking #AIWorkflow

Rourke Sefton-Minns

@rourke · creator

INSTAGRAM · 2025-09-12Source

2.2Klikes

91comments

Remix This

Recreate with Wan2.6

Make your own AI viral video

Prompt

GLOBAL LOCK: 
Subject: A Caucasian male in his late 20s with a dark beard and medium-length brown hair. 
Wardrobe: A cream-colored t-shirt and a tan "Vans" trucker hat with a red logo. 
Environment: A professional studio setup with a dark background featuring a glowing cyan/blue retro-futuristic perspective grid. 
Layout: A vertical 9:16 split-screen. The top 60% is a digital UI canvas (Krea AI interface). The bottom 40% is a talking-head overlay of the subject. 
Lighting: Soft three-point lighting on the subject; high-contrast digital glow on the UI. 
Color Grade: Saturated, clean, tech-focused palette with vibrant primary colors in the AI outputs. 
Speech: Natural, energetic UGC-style commentary, medium pace, crisp audio with slight room resonance.

[00:00–00:03]
Subject: Close-up of the talking head at the bottom, smiling and gesturing.
UI: Rapid montage of three split-screens: a green frog drawing becoming a 3D frog, a man in a hat becoming a realistic portrait, and an orange fish drawing becoming a photoreal goldfish.
Action: Subject points upward toward the UI.
Camera: Static for the overlay; fast cuts for the UI examples.
Speech: "This is one of the world's first real-time AI video creation tools..."

[00:03–00:10]
Subject: Subject gestures with his hands, explaining the process.
UI: A green canvas with a red circle on a thin green stem. As the mouse cursor moves the red circle, the bottom AI window shows a red flower blooming and shifting in real-time.
Action: Mouse cursor drags the red circle; the flower follows the movement perfectly.
Lighting: Bright, natural daylight feel in the AI flower window.
Speech: "...that allows you to move any element in your canvas and it will turn it into an AI video for you directly in front of your eyes."

[00:11–00:17]
Subject: Subject looks directly at the camera, nodding.
UI: A white background with a brown rectangular shape. The AI window shows a cup of tea. A red horizontal line is added, and the AI window reflects a tea-filled cup on a wooden surface.
Action: Adding geometric shapes to the canvas; AI updates the tea cup instantly.
Speech: "Now this is a brand new model from Krea and they've given me early access to show you exactly what's possible..."

[00:18–00:21]
Subject: Subject looks slightly to the side toward the UI.
UI: A photo of a living room is uploaded as a background. The tea cup is now composited into the living room scene in the AI window.
Action: Dragging an image file into the UI; AI blends the cup into the new environment.
Speech: "...you can upload images into the background as well to help sell realism in some scenes."

[00:22–00:27]
Subject: Subject holds hands up in a "wait" gesture.
UI: A black canvas with a teal rectangle and an orange circle. The AI window shows a glowing humanoid figure. A red triangle is added, and the AI window transforms it into a man sitting by a campfire.
Action: Abstract shapes are manipulated; the AI output shifts from a "glow" to a realistic campfire scene.
Speech: "Now don't get me wrong, there is a long way to go with this tech and it's not actually available yet but it will be very soon."

[00:28–00:31]
Subject: Subject points to the camera for the CTA.
UI: A blue and grey background with a yellow oval. The AI window shows a yellow Lamborghini sports car with headlights on.
Action: The yellow oval is moved; the car's perspective shifts in the AI window.
Text Overlay: "Follow for creative AI content" appears at the bottom of the UI.
Speech: "If you want to stay up to date with all the latest AI tech and trends, make sure you drop a follow."

NEGATIVE PROMPT: 
Visual: blurry face, distorted hands, flickering background grid, low resolution, watermark on creator, inconsistent hat logo, robotic movement, lag between mouse and AI output.
Speech: robotic voice, background noise, muffled audio, lip-sync delay, monotone delivery, harsh "S" sounds, clipping audio.

SPEECH PACK:
[00:00–00:03] "This is one of the world's first real-time AI video creation tools..."
TAKE_A: (Excited) This is one of the world's FIRST real-time AI video creation tools!
TAKE_B: (Informative) Check this out, it's one of the first real-time AI video tools ever made.
TAKE_C: (Fast) This is a world-first: real-time AI video creation.

[00:03–00:10] "...that allows you to move any element in your canvas and it will turn it into an AI video for you directly in front of your eyes."
TAKE_A: ...allowing you to move ANY element on your canvas and watch it turn into AI video right before your eyes!
TAKE_B: ...you just move things on the canvas and it generates the video instantly. It's magic.

[00:28–00:31] "If you want to stay up to date with all the latest AI tech and trends, make sure you drop a follow."
TAKE_A: Want more AI tech? Drop a follow to stay updated!
TAKE_B: Make sure you follow if you want to see the latest in creative AI.

How rourke Built This Krea Real-Time Workflow Video - and How to Recreate It

This case study analyzes a high-performing tech-review reel by creator @rourke, focusing on the "magic" of real-time AI video generation. The video showcases early access to Krea AI’s latest tool, utilizing a distinctive split-screen UI layout that contrasts simple geometric inputs with photorealistic AI outputs. The aesthetic is "Clean Tech UGC," featuring a talking-head overlay against a retro-futuristic grid background. By demonstrating the immediate transformation of a red circle into a blooming flower or a yellow oval into a speeding Lamborghini, the video taps into the "wow factor" of generative AI. It effectively bridges the gap between complex technology and creative play, making it highly shareable for both tech enthusiasts and casual creators looking for the "next big thing" in content production.

What You’re Seeing: Visual Breakdown

The video is structured as a fast-paced demonstration. The screen is divided into three main zones: a top header with the text "REALTIME AI VIDEO," a central split-screen showing the AI interface (top) and result (bottom), and a bottom overlay featuring the creator. The creator wears a casual "Vans" trucker hat and a cream t-shirt, positioned in a professional studio setup with soft, three-point lighting that separates him from the dark, grid-patterned background.

Shot-by-Shot Breakdown

Time Range	Visual Content	Shot Language	Lighting & Tone	Viewer Intent
00:00–00:03	Rapid montage of frog, man, and fish examples.	Fast cuts, split-screen.	Bright, high-contrast.	The Hook: Instant proof of concept.
00:03–00:10	A red circle on a stick is moved; a flower blooms in the AI window.	Screen recording + Talking head.	Clean, digital UI.	Demonstrate "Real-time" interaction.
00:11–00:17	A brown cup shape is placed on a table; tea appears.	Medium shot (overlay).	Warm, inviting AI colors.	Showcase versatility (objects).
00:18–00:21	Adding a background image to the tea cup scene.	UI interaction focus.	Natural room lighting in AI.	Tutorial value: How to "sell realism."
00:22–00:27	Abstract shapes turn into a man by a campfire.	Split-screen comparison.	Dramatic, low-key (firelight).	Awe: From "ugly" shapes to beauty.
00:28–00:31	Yellow oval becomes a car; CTA appears.	Final demo + Text overlay.	Vibrant, saturated yellow.	Retention & Conversion (Follow).

Why It Went Viral: The Magic of Real-Time

The primary driver of this video's success is the "Instant Gratification" loop. In a world where AI video usually takes minutes to render, seeing a flower change shape exactly as the mouse moves triggers a biological "Aha!" moment. The选题 (topic selection) targets the intersection of curiosity and utility. It’s not just a cool trick; it’s a glimpse into the future of professional workflows. The creator uses "Early Access" as a social currency, positioning himself as an authority while providing exclusive value to his audience.

From a platform perspective, the split-screen format is a retention machine. The viewer's eyes are constantly moving between the "Input" (top) and the "Output" (bottom) to verify the AI's accuracy. This increases visual engagement and prevents the "scroll-past" reflex. The 0–3 second hook uses a rapid-fire montage of three different subjects, signaling to the algorithm that this video has high information density, which typically leads to higher completion rates and shares to "Saved" folders for future reference.

5 Testable Viral Hypotheses

The "Input-Output Contrast" Hypothesis: Showing a "bad" or "simple" drawing next to a "perfect" AI result creates a satisfying transformation that viewers love to watch repeatedly.
The "Real-Time Proof" Hypothesis: Including the mouse cursor and UI elements proves the video isn't a fake edit, building trust and increasing the "wow" factor.
The "Early Access" Authority: Mentioning a specific tool (Krea AI) and "early access" creates a sense of exclusivity, encouraging users to comment "How do I get this?" which boosts engagement.
The "Utility vs. Play" Balance: By showing a flower (art) and then a car (commercial), the video appeals to both hobbyists and professional creators.
The "Talking Head" Trust Factor: Having the creator visible at the bottom humanizes the tech review, making the information feel like a recommendation from a friend rather than an ad.

How to Recreate: From 0 to 1

Step 1: Tool Selection & Setup

Identify a "Real-Time" or "Interactive" AI tool (e.g., Krea AI, Leonardo Canvas, or ComfyUI Real-time). Ensure you have a screen recording setup that captures both your mouse movements and the AI's response without lag.

Step 2: The "Talking Head" Environment

Set up your camera (iPhone is fine) with a clean background. Use a grid overlay or a simple LED-lit shelf. Wear a hat or distinct accessory to maintain character consistency across your reels.

Step 3: Scripting the "Real-Time" Hook

Your first 3 seconds must show at least 3 different transformations. Don't explain; just show. Use a voiceover that says: "This is the first time we've seen AI do [X] in real-time."

Step 4: Recording the Demo

Record 4-5 distinct examples. Start with something simple (a circle) and move to something complex (a character or vehicle). Actionable Tip: Move the shapes slowly so the AI has time to "morph" on screen—this is the most satisfying part for the viewer.

Step 5: Editing the Split-Screen

In CapCut or Premiere, stack your screen recording on top and your talking head at the bottom. Use a "Mask" or "Crop" tool to create the clean horizontal split seen in the video. Add a header text like "REALTIME AI VIDEO" to provide context for silent scrollers.

Step 6: Sound Design

Use a "Tech-Ambient" or "Phonk" background track at low volume. Ensure your voiceover is crisp. Add subtle "pop" or "whoosh" sound effects when the AI output changes significantly.

Step 7: Captions & Overlays

Add "Available publicly soon!" or tool names as text overlays during the video. This answers the viewer's most likely question before they even ask it.

Step 8: Publishing Strategy

Post to Instagram and TikTok simultaneously. Use the "Collaborator" feature if you are reviewing a specific tool to tap into their audience.

Growth Playbook: Hooks & Captions

3 Opening Hook Lines

"AI video just hit 0ms latency and it’s terrifying. 🤯"
"Stop waiting for renders. This tool generates video as you draw."
"I got early access to the 'Photoshop' of AI video. Look at this..."

4 Caption Templates

The Hype: "Real-time AI video is finally here. 🤯 I’ve been testing @krea_ai and the speed is insane. Which example was your favorite? 👇"
The Tutorial: "How to use shapes to control AI. 🎨 1. Draw a circle. 2. Watch it become a flower. 3. Add a background. The future of design is interactive. Save this for later! 💾"
The Skeptic/Discussion: "Is real-time AI a gimmick or a game changer? 🧐 It’s fun to play with, but can it be used for real client work? Let me know your thoughts in the comments. 💬"
The "Waitlist" CTA: "The waitlist for this is going to be huge. 🚀 Watch me turn a yellow oval into a supercar in 2 seconds. Follow for the link when it drops! 🔥"

Hashtag Strategy

Broad: #AI #ArtificialIntelligence #TechNews #FutureTech
Mid-Tier: #AIVideo #GenerativeAI #KreaAI #ContentCreatorTools
Niche: #RealTimeAI #AIWorkflow #MotionDesignAI #IndieCreator

Frequently Asked Questions

What tool is being used in this video?

The creator is using Krea AI's real-time video generation tool.

How do I get the split-screen look?

Use a video editor like CapCut to stack a screen recording over your camera footage and apply a crop.

Why does the AI look "jittery" sometimes?

Real-time AI often has lower temporal consistency because it's generating frames on the fly without looking at previous ones.

Is this available for mobile?

Most of these tools are web-based and work best on a desktop browser with a strong GPU.

How can I avoid the "AI look"?

Use the "Image Prompt" feature (as shown at 0:18) to give the AI a high-quality reference for lighting and texture.

Do I need to disclose this is AI?

Yes, most platforms now require an "AI-Generated" label for photorealistic content to maintain transparency.