Google’s new Nano Banana Pro is otherworldly. Here are 3 super useful things you can do with this, that you could not do before with any other AI. First off, the character consistency is crazy. For the first time, you can actually build a world, with any person in any place, in any style, that actually looks good. You can change their look, their clothes, their expression, the angle of the cameras, their movement, the lighting in the environment. Any design lever you wanna pull, you can pull. And the character you build will stay perfectly consistent the entire time. You can do this with a single photo of you, a celebrity, or any generated character. It’s all super easy to do in @higgsfield.ai . All of these visuals came directly out of Higgsfield without any editing. Now the second thing you can do with this model is even more interesting. And I haven’t seen anybody cover this. Nano Banana Pro is incredible with text. So you can drop in an entire PDF and it will revisualize it into a diagram or infographic. And this is gonna completely change learning. Full textbooks will get rerendered into educational picture books. Text-heavy power point slides are gonna be visualized instantly. Imagine anything with lots of text that is boring to look at…menus, brochures, even handwritten notes…can all be turned into images whenever you want. Now the third thing is the most underrated. And that’s the integration with Google. Because Nano Banana Pro is a Google product, it has unlimited access to Google’s entire ecosystem. Maps, YouTube, and tons of real-world data. If you plug in specific geographic coordinates, let’s say Central Park, South East corner in 1979, it can generate that image, from that spot, in that era, perfectly. Now we are just scratching the surface on what this is gonna be able to do. But the reality is…the creative world is changing rapidly. If you wanna try out Nano Banana Pro, Higgsfield’s the easiest way. And if you wanna stay updated on everything AI, make sure to follow @kallaway #ai #artificialintelligence #tech #technology #google #nanobananapro #design

Kallaway

@kallaway · Entrepreneur

INSTAGRAM · 2025-11-24Source

1.3Klikes

17comments

Remix This

Prompt

GLOBAL LOCK:
The video follows a split-screen format. The bottom half features a consistent male creator: Caucasian, mid-20s, wearing a black Siegelman Stable baseball cap and a black "Suspicious Antwerp" hoodie with a white/red crest. The setting is a warm, low-light studio with a palm tree silhouette in the background. The top half features various AI-generated scenes. The overall style is high-energy tech-editorial with fast cuts and bold text overlays.

[00:00–00:03]
TOP: Rapid sequence of high-quality AI renders: 1) Donald Trump in a dark suit eating a peeled banana in a lush green jungle. 2) Keanu Reeves as John Wick in a CS:GO-style desert map holding a rifle. 3) A female Fortnite character in a galaxy-patterned bodysuit holding a microphone.
BOTTOM: Creator talking enthusiastically, gesturing with hands.
SPEECH: "Google's new Nano Banana Pro just dropped."
SYNC: High lip-sync strictness.

[00:04–00:10]
TOP: Cinematic close-up of a young girl's face with intense blue eyes and freckles, followed by a shot of a transparent, glass-like vintage car in a field of tall grass.
BOTTOM: Creator points upward. Text "otherworldly" appears in white script font.
SPEECH: "Nano Banana Pro. It's their brand new AI image model. And this thing is otherworldly."

[00:11–00:27]
TOP: Character consistency demo. Michael Jackson in a sequined jacket and fedora, side-by-side with a reference photo. Then, actor Lee Jung-jae in a green tracksuit standing by a vintage green car in a suburban street. Then, Pedro Pascal in a denim outfit doing pushups in a white studio.
BOTTOM: Creator explains the feature, counting on his fingers.
SPEECH: "First off, the character consistency is crazy. For the first time, you can actually build a world with any person, in any style, that actually looks good."

[00:28–00:36]
TOP: A "flash-photography" style party photo. A young man (the creator) holding a red solo cup, smiling next to Leonardo DiCaprio. Then, the same man in a banana suit next to DiCaprio.
BOTTOM: Creator points to the top screen.
SPEECH: "You can do this with a single photo of you, a celeb, or any generated character. It's all super easy to do in Higgsfield."

[00:37–00:55]
TOP: Text and diagram demo. A green "Café Cuillère d'Or" sign. Then, complex technical diagrams of a robot, a water cycle, and a physics equation on a chalkboard. Then, a restaurant menu with food photos.
BOTTOM: Creator gestures broadly.
SPEECH: "Now the second thing you can do with this model is even more interesting. Nano Banana Pro is incredible with text. You can drop in an entire PDF and revisualize it as a diagram or an infographic."

[00:56–01:15]
TOP: Google ecosystem integration. Icons for Gmail, Drive, and Maps. Then, a screen recording of Google Maps zooming into NYC. Finally, a high-quality AI render of a 1979 NYC street corner with vintage yellow cabs and people in period clothing.
BOTTOM: Creator looks impressed, nodding.
SPEECH: "And the third thing is the integration with Google. It has unlimited access to Google's entire ecosystem. Maps, YouTube, Gmail... it can render that exact visual at that exact time perfectly."

[01:16–01:29]
TOP: A melting soft-serve ice cream cone next to a vintage street clock. Then, a manga-style comic page showing two characters fighting.
BOTTOM: Creator smiles, gives a thumbs up, and points to the "Follow" button area.
SPEECH: "We're just scratching the surface. The creative world is changing rapidly. If you want to try it, Higgsfield is the easiest way. Stay updated, make sure to follow."

NEGATIVE PROMPT:
Visual: blurry faces, distorted hands, flickering backgrounds, inconsistent clothing colors, low-resolution textures, AI "hallucinations" in text, robotic movements.
Speech: monotone delivery, robotic cadence, background noise, muffled audio, poor lip-sync, unnatural pauses.

SPEECH PACK:
[00:00-00:10]
TAKE_A: "Google’s new Nano Banana Pro just dropped. It’s their brand new AI image model. And this thing is... otherworldly." (Excited, fast pace)
TAKE_B: "So, Google just released Nano Banana Pro. A brand new AI model that is honestly... otherworldly." (Casual, conversational)

[00:11-00:27]
TAKE_A: "First off, character consistency is crazy. You can build a world with any person, in any style, and it actually looks good." (Emphasize "crazy" and "any")
TAKE_B: "The consistency here is next level. Any person, any style, any place. It just works." (Punchy, short sentences)

[00:59-01:15]
TAKE_A: "The Google integration is the most underrated part. It uses real-world data to render specific locations and times perfectly." (Informative, authoritative)
TAKE_B: "Think about the Google data. Maps, YouTube, everything. It renders historical moments exactly how they looked." (Awe-struck tone)

How kallaway Made This Nano Banana Pro Tutorial AI Video

This case study analyzes @kallaway's high-impact educational video featuring the "Nano Banana Pro" AI model. The video utilizes a split-screen "Proof + Persona" format: the top half showcases mind-bending AI-generated visuals (cinematic portraits, historical recreations, and complex diagrams), while the bottom half features the creator in a warm, low-light studio setting. With a warm, tech-noir aesthetic, the video achieves high retention through rapid-fire visual evidence of character consistency and real-world data integration. It’s a masterclass in "Newsjacking" AI developments while maintaining a consistent personal brand through wardrobe (black hoodie/cap) and a signature editing rhythm.

What You’re Seeing: A Detailed Breakdown

The video is a vertical 9:16 edit. The top 50% is a dynamic gallery of AI outputs—ranging from Donald Trump eating a banana in a jungle to Pedro Pascal doing pushups in a denim outfit. The bottom 50% is a talking-head shot of @kallaway. The lighting is motivated by practicals in the background (warm lamps, neon-style shelf lights), creating a "pro-sumer" home studio vibe. The color grade is moody with deep blacks and warm skin tones. Text overlays are minimalist, using a bold sans-serif font that "pops" to emphasize key terms like "otherworldly" or "consistency."

Shot-by-Shot Breakdown

Time Range	Visual Content	Shot Language	Lighting & Tone	Viewer Intent
00:00–00:03	Trump eating banana / Keanu in CS:GO / Fortnite skin.	Split-screen; rapid cuts.	High-contrast AI renders.	The "WTF" Hook: Absurd but high-quality visuals.
00:04–00:10	Creator talking; "otherworldly" text overlay.	MCU (Medium Close Up).	Warm studio, shallow DOF.	Establish authority and set the agenda.
00:11–00:27	Michael Jackson / Pedro Pascal consistency demo.	Side-by-side AI comparisons.	Cinematic editorial lighting.	Proof of Value: Solving the "consistency" pain point.
00:28–00:36	Creator at a party with Leo DiCaprio.	UGC-style "fake" photo.	Flash photography aesthetic.	Relatability: "You can do this too."
00:41–00:55	Text-to-diagram / Menu / PDF examples.	Clean graphic overlays.	Bright, legible, informative.	Utility: Showing AI for productivity, not just art.
00:59–01:15	Google Maps / Historical NYC street view.	Screen recording + AI render.	Vintage film look (1979).	The "Magic" Moment: Merging data with imagination.

Why It Went Viral: The "Banana" Effect

The Power of "Absurd Authority"

The选题 (topic selection) is brilliant because it combines absurdity with high-utility tech news. Starting with a world leader eating a banana (biological instinct + curiosity) immediately stops the scroll. However, the video quickly pivots to solving the #1 problem in AI video: Character Consistency. By using recognizable celebrities like Pedro Pascal and Michael Jackson, the creator leverages the "Familiarity Heuristic"—users already know what these people look like, so the AI's accuracy is immediately verifiable.

The "Nano Banana" Hook

The use of a quirky name like "Nano Banana Pro" acts as a pattern interrupt. In a sea of "GPT-5" and "Sora" news, a "Banana" model sounds like a joke, which lowers the viewer's guard and increases curiosity. This "mild controversy" or "oddity" is a classic psychological trigger that encourages comments asking, "Wait, is this real?"

Platform Perspective: The Retention Loop

From a platform perspective, this video is a retention monster. The split-screen format provides two streams of information simultaneously. If the viewer gets bored of the creator's face, their eyes drift to the top-half AI visuals. The loop effect is achieved by ending on a high-energy call to action ("make sure to follow"), which feels like a natural conclusion to a high-value session. The "Save" rate is likely high because the video acts as a "Prompt Engineering" reference for tools like Higgsfield.

5 Testable Viral Hypotheses

The Split-Screen Synergy: If you show the result (AI video) and the process/commentary (creator) at the same time, watch time increases by 20% compared to full-screen talking heads.
Celebrity Benchmarking: Using a globally recognized face (e.g., Pedro Pascal) to demonstrate a tech feature provides instant "proof of quality" that generic AI characters cannot achieve.
The "Absurd Hook" Lead: Starting with a high-quality but nonsensical image (Trump + Banana) triggers a "What happens next?" reflex more effectively than a standard "Hey guys" intro.

Utility Over Art:

text

Historical Time-Travel: Recreating a specific year/location (NYC 1979) via Google Maps coordinates creates an emotional "wow" factor that drives shares.

How to Recreate: From 0 to 1

Step 1: Topic Selection & Positioning

Choose a "New Tech" angle. This format works best for AI, software updates, or "life hacks." Your account should position itself as the "Curator of the Future."

Step 2: The "Absurd" Hook Asset

Generate 3-5 high-quality, slightly weird AI images using Midjourney or DALL-E 3. Think: "Historical figures in modern settings" or "World leaders doing mundane things."

Step 3: Character Consistency Setup

Use a tool like Higgsfield or Midjourney's --cref (Character Reference) feature. Pick one character and generate them in 5 different outfits and 3 different lighting setups to prove the "consistency" point.

Step 4: Filming the Persona

Film your talking-head segment in a low-light environment. Use a single key light at a 45-degree angle to create depth. Wear a neutral, branded outfit (hoodie/cap) to keep the focus on the content.

Step 5: The Split-Screen Edit

In CapCut, use the "Overlay" feature. Place your talking head on the bottom and your AI assets on the top. Ensure the cuts in the top half happen every 1.5 to 2 seconds to match the "fast-paced" tech vibe.

Step 6: Text Overlay Strategy

Add "Dynamic Captions." Highlight only the power words (e.g., "Consistency," "Text," "Google"). Use a high-contrast color like yellow or white on a dark background.

Step 7: BGM Selection

Search for "Phonk" or "Tech House" tracks with a steady, driving beat. Lower the volume to 10-15% so your voice remains the primary focus.

Step 8: The "Loop" CTA

End your script with a question or a "Follow for more" that starts just as the last visual finishes, encouraging the viewer to watch the loop again to catch details they missed.

Growth Playbook: Distribution & Scaling

3 Ready-to-Use Opening Hooks

"The new [Model Name] just dropped and it's actually terrifying."
"I found the secret to perfect AI character consistency, and it’s not what you think."
"Google just integrated AI into [App Name] and it changes everything."

4 Caption Templates

The "Value Drop" Template:
Google’s new Nano Banana Pro is otherworldly. 🍌 Here are 3 things you can do now that were impossible last week. [Value 1], [Value 2], [Value 3]. Which one are you trying first? 👇 #AI #TechNews

The "How-To" Template:
Stop struggling with AI consistency. 🛑 This new workflow using [Tool] lets you build entire worlds with one character. Check the video for the prompt logic. Save this for your next project! #AITutorial #ContentCreator

Hashtag Strategy

Broad: #AI #Technology #FutureTech #Innovation
Mid-Tier: #AIArt #VideoEditing #ContentStrategy #Higgsfield
Niche: #CharacterConsistency #AIPrompting #NanoBanana #KallawayStyle

Frequently Asked Questions

What tools make the AI video look this consistent?

The creator specifically mentions Higgsfield for character consistency and video generation.

How do I get the "split-screen" look?

Use the "Mask" or "Overlay" tool in CapCut or Premiere Pro to stack two video tracks.

Is "Nano Banana Pro" a real Google model?

It appears to be a creative/fictional name used as a hook for demonstrating Gemini/Higgsfield capabilities.

How can I avoid the "AI look" (flickering)?

Use lower "motion buckets" in your generation settings and keep the camera movement static in the prompt.

Why is the creator wearing a hat and hoodie?

This creates a "Uniform" or "Persona" that makes the creator instantly recognizable in the feed.