0:00 / 0:00

Ylenia Fedi

@yleintech

INSTAGRAM · 2026-03-11Source

134likes

251comments

Remix This

Recreate with Kling 3

Make your own AI viral video

Prompt

GLOBAL LOCK: 
Subject: The Empire State Building in New York City. 
Identity: Photorealistic architectural landmark, Art Deco style, limestone and steel materials.
Environment: Mid-town Manhattan city grid, dense skyscrapers, clear sky.
Lighting: Golden hour sunset, warm orange light hitting the building's west face, long shadows across the city.
Color Grade: Cinematic, high contrast, vibrant warm tones, deep blue shadows.
Camera: High-altitude drone perspective, smooth cinematic movement.
Speech: Female narrator, mid-20s, enthusiastic and clear instructional tone, professional studio microphone quality.

[00:00–00:05]
Visual: A cinematic aerial zoom-in towards the upper half of the Empire State Building. The camera moves smoothly forward. The city below is detailed with moving cars and soft shadows.
Text Overlay: "HOW TO MAKE THIS" in bold white sans-serif font, followed by "into a timelapse" in smaller text.
Speech: "You can turn one simple photo of a building into a full construction timelapse in just a few clicks."
Sync: Cut to next shot on the word "clicks."

[00:05–00:10]
Visual: Screen recording of a desktop browser. The cursor navigates to the Higgsfield.ai website. It clicks on a card labeled "Nano Banana Pro."
Speech: "First, take a photo of any building. Then go to Higgsfield and open Nano Banana Pro."
Sync: UI interaction matches the narration of tool names.

[00:10–00:14]
Visual: A top-down aerial photo of a New York City block. A bright red glowing square highlights a rectangular foundation pit where a building should be.
Speech: "Upload your photo and paste the prompt. It will generate an image showing the building's foundation stage."
Sync: The red highlight appears as the narrator says "foundation stage."

[00:14–00:21]
Visual: Screen recording of the Kling AI interface. Two image slots are filled: the foundation pit on the left and the finished Empire State Building on the right. A JSON-style prompt is pasted into the text box. The cursor clicks a green "Generate" button.
Speech: "Next, open Kling AI and choose Kling 3.0 Video. Upload both foundation and original image, paste the second prompt and hit generate."
Sync: The click on "Generate" happens exactly at 00:21.

[00:21–00:24]
Visual: The final AI-generated video. From a fixed aerial camera angle, the Empire State Building rapidly grows upward from the foundation. Scaffolding flickers, floors stack instantly, and the spire completes at the top. The lighting remains consistent golden hour.
Speech: "In a few minutes, your video will be ready."
Sync: The building reaches full height as the sentence ends.

[00:24–00:27]
Visual: A medium close-up of a woman with long dark hair and a white ribbed top. She is in a studio with a blurred background featuring purple and pink neon tube lights. She gestures with her hands while speaking.
Text Overlay: "Comment Time" in large yellow and white text.
Speech: "Comment 'Time' and I'll send you the exact prompt I used."
Sync: Lip-sync is high strictness; the video ends on her smile.

NEGATIVE PROMPT: 
Visual: blurry textures, distorted architecture, flickering lighting, inconsistent shadows, warped windows, floating debris, sudden camera jumps, low resolution, watermarks, distorted human features in the outro.
Speech: robotic voice, background noise, muffled audio, stuttering, unnatural pauses, misaligned lip-sync, harsh "S" sounds, clipping audio.

SPEECH PACK:
[00:00-00:05]
TAKE_A: "You can turn one simple photo of a building into a full construction timelapse in just a few clicks." (Energetic, fast pace)
TAKE_B: "Turn any single photo of a building into a professional construction timelapse with just a few clicks." (Authoritative)
TAKE_C: "Imagine turning one photo into a full construction timelapse. It only takes a few clicks." (Mysterious/Intriguing)

[00:24-00:27]
TAKE_A: "Comment 'Time' and I'll send you the exact prompt I used." (Friendly, direct)
TAKE_B: "Want the prompt? Just comment 'Time' below and I'll send it over." (Casual)
TAKE_C: "Comment 'Time' right now to get the exact prompt for this workflow." (Urgent)

PROSODY MARKUP:
"You can turn **one simple photo**... into a **full** construction timelapse... in just a few **clicks**."
"Comment **Time**... and I'll **send** you the exact prompt I used."

How yleintech Made This AI Building Construction Timelapse Tutorial — and How to Recreate It

This viral tutorial by @yleintech demonstrates a high-impact "magic trick" for content creators: transforming a single static photograph of a landmark into a professional-grade construction timelapse using AI. The video features a cinematic aerial view of the Empire State Building in New York City, bathed in warm golden hour light. By leveraging advanced AI tools like Higgsfield and Kling AI 3.0, the creator shows how to "reverse engineer" a building's existence, starting from a foundation pit and growing into a skyscraper in seconds. This case is a perfect example of "Utility-Aesthetic" content—it provides a clear, actionable tutorial while maintaining a high-end, cinematic visual style that stops the scroll. The use of a recognizable global icon (NYC skyline) combined with a "comment for the prompt" call-to-action (CTA) creates a powerful engine for engagement and reach.

What You’re Seeing

The video is a hybrid of cinematic drone-style footage, screen-recorded software tutorials, and a direct-to-camera "talking head" conclusion. It begins with a sweeping zoom into the Empire State Building, immediately establishing a high-production-value feel. The color palette is dominated by warm oranges and deep blues, typical of a sunset in Manhattan. As the tutorial progresses, we see the user interface of AI platforms, specifically highlighting the Nano Banana Pro model on Higgsfield and the Kling 3.0 video generator. The most striking visual is the "foundation stage" image—a red-tinted architectural footprint that serves as the starting point for the AI's temporal generation. The video concludes with the creator in a studio setting, using purple and pink neon practical lights to create a "tech-expert" persona, reinforcing her authority in the AI space.

Shot-by-Shot Breakdown

Time Range	Visual Content	Shot Language	Lighting & Tone	Viewer Intent
00:00–00:05	Aerial zoom into Empire State Building with "HOW TO MAKE THIS" text.	Cinematic Drone / Zoom-in	Golden hour, high contrast	Hook: Show the "impossible" result immediately.
00:05–00:10	Screen recording of Higgsfield website and Nano Banana Pro tool.	UI Capture / Screen Recording	Clean, digital interface	Instruction: Establish the first tool needed.
00:10–00:14	Image of a city block with a red highlighted foundation pit.	Static Image with Overlay	Daylight, architectural	Clarity: Show the "secret" intermediate step.
00:14–00:21	Kling AI interface showing image uploads and JSON prompt pasting.	UI Capture / Screen Recording	Dark mode UI, tech-focused	Process: Demonstrate the actual "work" being done.
00:21–00:24	Final timelapse: Building grows from ground to sky in NYC.	AI Generated Timelapse	Consistent with Shot 1	Payoff: Prove the method works with a "wow" visual.
00:24–00:27	Creator @yleintech speaks to camera, "Comment Time" text.	Medium Close-up (MCU)	Studio, Pink/Purple Neon	CTA: Convert viewers into engaged followers.

Why It Went Viral

The success of this video lies in its "Information Gap" strategy. By showing a complex result (a construction timelapse) and then revealing that it only takes "a few clicks," the creator triggers intense curiosity. The choice of the Empire State Building is strategic; it’s a globally recognized symbol of ambition and architecture, making the video relevant to a massive, international audience. Psychologically, humans are fascinated by growth and construction (the "SimCity" effect), and seeing a massive structure assemble itself in seconds taps into a deep-seated biological interest in environmental transformation.

From a platform perspective, the video is engineered for Watch Time and Saves. The fast-paced editing (cuts every 3-5 seconds) prevents boredom, while the technical nature of the tutorial encourages users to "Save" the video to reference the tool names later. The "Comment 'Time' for the prompt" tactic is a masterclass in algorithm manipulation. It forces users to engage, which signals to Instagram that the content is highly valuable, triggering a wider distribution to the Explore page. Furthermore, the creator’s professional studio setup at the end builds trust and personal branding, making viewers more likely to follow for future "tech secrets."

5 Testable Viral Hypotheses

The "Iconic Landmark" Hook: Using a world-famous building (Empire State, Eiffel Tower) increases the initial click-through rate because the subject is instantly recognizable.
The "Reverse-Engineering" Curiosity: Showing the foundation stage (the "before") creates a mental itch that the viewer must see resolved by the "after."
The "JSON Prompt" Authority: Showing a snippet of code or a complex-looking JSON prompt makes the creator look like an expert, increasing the perceived value of the "free" gift.
The "Tool Stacking" Value: Mentioning two specific, niche tools (Higgsfield + Kling) makes the tutorial feel more exclusive and "pro" than a generic "use ChatGPT" video.

The "Low Friction" CTA:

How to Recreate (Step-by-Step)

Subject Selection: Choose a high-resolution photo of a prominent building. Ensure the building has clear edges and a distinct architectural style.
Establish the "Before": Use an image editor or an AI tool like Higgsfield (Nano Banana Pro) to generate an image of the same location but with the building removed, showing only the foundation or a flat construction site.
Maintain Consistency: Ensure the lighting and camera angle in your "foundation" image perfectly match your "finished" photo. Use the original photo as a reference.
Video Generation Tool: Open Kling AI (Version 3.0). This specific version is known for its "Image-to-Video" consistency and temporal control.
The "Keyframe" Method: Upload your foundation image as the "Start Frame" and your original building photo as the "End Frame."
Prompting for Motion: Use a prompt that describes upward growth, scaffolding appearing, and glass panels fitting into place. (See the Master Prompt section below).
Screen Recording: Record your process in the browser. Use a tool like Loom or built-in OS recorders. Crop the video to 9:16 for mobile viewing.
The "Expert" Outro: Film yourself in a well-lit environment (use colored LED bars for a "tech" vibe) and deliver a clear CTA that promises more value in exchange for engagement.

Growth Playbook

3 Opening Hook Lines

"Stop scrolling! Here is how to build a skyscraper in 5 seconds using AI."
"The secret to those viral construction timelapses? It’s actually just two tools."
"You don't need a drone or a year of filming to make this. Just one photo."

4 Caption Templates

Option 1 (The Direct Value):
Turn any photo into a construction timelapse! 🏗️
I used Kling 3.0 and Higgsfield to make this happen.
Want the exact workflow?
👇 Comment "PROMPT" and I'll DM it to you!

Hashtag Strategy

Broad (Reach): #AI #Tech #Innovation #Architecture #NewYorkCity
Mid-Tier (Niche): #AIVideo #KlingAI #ContentCreatorTips #VideoEditing #DigitalArt
Long-Tail (Specific): #ConstructionTimelapseAI #HiggsfieldTutorial #EmpireStateBuildingAI #IndieCreatorTools

Frequently Asked Questions

What tools make it look the most similar?

Kling AI 3.0 for the video generation and Higgsfield for the initial foundation image generation.

What are the 3 most important words in the prompt?

"Construction," "Timelapse," and "Growth-sequence."

Why does the generated face look inconsistent?

In this specific video, the building is the focus; for faces, use a "Character Reference" (Cref) tool in Midjourney or Kling.

How can I avoid making it look like AI?

Use a high-quality real photograph as the "End Frame" to ground the AI's output in reality.

Is it easier to go viral on Instagram or TikTok with this?

Instagram, as the "aesthetic tutorial" niche is currently booming there with high save rates.