AmirMušić Profile picture
I help creators & businesses grow in the AI era. Ex: @warnermusic
2 subscribers
Dec 19, 2025 4 tweets 3 min read
Nano Banana auto-prompt

Unexpected high-end product concepts

Prompt below👇 Image Just change [BRAND NAME] and enjoy the endless, high-end ideas visualized as it was a concept magazine

👉 Prompt:

[BRAND NAME]:
A high-end, glossy concept art magazine editorial photograph of a unique, unexpected functional object conceptualized and designed by the brand.

**1. The Concept & Object (AI Invention):**
Based on the design philosophy, heritage, and material vocabulary of the specified brand, the AI must invent a novel utility product (NOT standard clothing, shoes, or bags). Examples could be home goods, tech accessories, tools, or sporting equipment, reinterpretated through the brand's lens. The object should feel sculptural yet functional.

**2. Materials & Details (Hyper-Premium):**
The object is constructed from ultra-premium, highly tactile materials characteristic of the brand (e.g., patinated exotic leathers, brushed aerospace-grade titanium, sculpted matte ceramics, molded carbon fiber, or technical high-fashion textiles). Every detail is hyper-realistic: visible stitching, microscopic material grain, precision engravings, and complex texture contrasts.

**3. Photography & Lighting (Cinematic Studio):**
Shot on a medium format Phase One camera with a 100mm macro lens. Extremely shallow depth of field, with sharp focus on the hero details of the object and a creamy, smooth bokeh background. The lighting is sophisticated studio softbox lighting: gentle, enveloping fill light with precise rim lighting to accentuate contours and material textures.

**4. Environment:**
A seamless, impeccably clean studio cyclorama background in a pure, ultra-light pastel tone (e.g., desaturated mint, pale blush, or off-white), free of shadows.

**5. Layout & UI Elements (Strict Placement):**
- **Bottom Right Corner:** A small, understated, monochrome gray logo of the brand.
- **Bottom Left Corner:** Small, minimalist monochrome gray text describing the invented product. The font style looks like Manrope Regular with very tight tracking (kerning) and balanced line spacing. Example format: "CONCEPT STUDY: [AI inserts invented product name]. MATERIAL: [AI inserts main materials]. SS25."Image
Dec 12, 2025 5 tweets 3 min read
Gemini x Nano Banana system prompt

Hire a $15,000-worth, viral product designer

Prompt & tutorial below: Image Step 1
Paste this system prompt in any LLM:

SYSTEM PROMPT: The Viral Product Architect
ROLE:You are a world-class Concept Art Director & Product Designer with 20+ years of experience at agencies like Pentagram, IDEO, and DONDA. Your specialty is "Speculative Design" and "Hype-culture artifacts". You create viral, thumb-stopping product concepts that mix High Fashion, Tech, and Utility in unexpected ways.

OBJECTIVE:The user will provide ONLY a [BRAND NAME]. Your task is to invent a hyper-realistic, unexpected, and visually stunning physical product concept for this brand. It must be something they don't sell, but could strictly hypothetically design if they went crazy.

CORE ALGORITHM (THE "FRESHNESS" FORMULA):

Analyze Brand DNA: Extract the brand's iconic materials, colors, and textures (e.g., Adidas = Primeknit/Boost foam; Rimowa = Aluminum grooves; Jacquemus = Tiny, leather).

Select Unexpected Object: Choose a mundane or tech object completely unrelated to the brand (e.g., Toothbrush, Inhaler, Drone, Smart Ring, VR Headset, Coffee Press).

Apply Texture Mapping: Wrap the "Unexpected Object" in the "Brand DNA". Crucial: Use materials typically reserved for fashion/furniture on tech items (like the silver puffer watch).

Viral Check: Ask yourself: "Would Highsnobiety or Hypebeast repost this?" If no, make it weirder.

VISUAL STYLE GUIDELINES (IMMUTABLE):

Vibe: Ultra Hi-End Product Design, Conceptual, Expensive, Tactile.

Environment: Minimalist light studio, infinity white/light grey background.

Lighting: Softbox studio lighting, clean reflections, high-key, ambient occlusion.

Camera: Telephoto lens (50mm-85mm), shallow depth of field, sharp focus on texture details.

OUTPUT FORMAT:You must output a SINGLE, OPTIMIZED IMAGE PROMPT (English) tailored for Google Imagen/Midjourney. Do not write an explanation. Just the prompt.

PROMPT STRUCTURE TO GENERATE: [Subject Description with Brand Materials] + [Context/Action] + [Specific Brand Details/Logo placement] + [Lighting & Environment] + [Technical 3D Specs]Image
Dec 4, 2025 8 tweets 4 min read
5-step "Style factory" workflow in Nano Banana:

1️⃣ Open Gemini app or LTX . studio
2️⃣ Upload your photo
3️⃣ Enter the base prompt (check 1st comment )
4️⃣ Change clothes & colors in [BRACKETS]
5️⃣ Hit Generate & repeat for different styles

Steal the prompt + action plan 👇 Image 👉 Take this base prompt:
Change text in [square brackets] - use the image below as a hint

🔽🔽🔽

Main model (use image input with the face 100% accurate).

Wearing:
[detailed clothing description - pieces, color scheme, style, layers, fabrics].

Paired with footwear:
[type and details of shoes; include relevant accessories like socks, hats, bags, jewelry, etc.].

Setting:
studio backdrop in soft [describe color/tone/texture - e.g., pastel, urban, concrete, natural, saturated].

Lighting:
soft cinematic studio lighting

Style:
[define style - for example: fashion editorial, streetwear, minimalism, preppy, futuristic, luxury, sporty, etc.].

Composition:
model [describe pose - seated/standing/reclined, relaxed/confident/youthful] on [describe object if relevant - bench, chair, wall], with [accessories/pose details - e.g., accessories visible, hands in pocket, dramatic posture].

💡 (prompt inspired by @ecommartinez)Image
Dec 1, 2025 4 tweets 2 min read
Nano Banana prompt

Swiss design geometric logo stylization

Prompt below👇 Image 👉 Prompt:
(change [BRAND] and [COLOR])

3D embossed glossy contour render of center-aligned [BRAND] on a flat surface, perfectly centered composition with ample negative space surrounding the object for a premium minimalist aesthetic. Monochromatic [COLOR] palette with soft tonal gradients. The object is defined by a raised, smooth, liquid-like glass bezel or chrome rim, creating a blind emboss effect where the interior matches the background. Matte surface finish with fine film grain or noise texture overlay. Soft diffuse lighting, strong specular highlights on the rounded edges, top-down view.
Nov 21, 2025 10 tweets 4 min read
5 ideas for Nano Banana you want to test

- 5 ideas,
- 5 prompts
- endless use cases

Steal them now, and save this thread for later 👇 Image 1/

Turn anything into a technical drawing
Inspired by @icreatelife & @CaptainHaHaa

Prompt:
Create a technical drawing of this photo, make it [visionary]

Additional prompt:
now make a different angle and highlight a different features Image
Nov 20, 2025 4 tweets 2 min read
Nano Banana Pro test: 8 images input
+ prompts 👇

Google's top image model is out

And I'm stress-testing it:
• 8 images input
• 1 prompt
• Task: to create a Hollywood-grade poster

The result is cool, but.. we have a PROBLEM.

Don't start until you read this 👇Image More image inputs = less consistency

I ran the model 25 times to achieve the result

The model was
• changing characters
• losing characters
• added unwanted characters

I spent around $8 on @fal to achieve the result
(see the video)

There's more 👇
Nov 2, 2025 9 tweets 5 min read
Brands pay agencies $5,000+ for videos.

Today, I'm giving you the entire agency for free.
From pro showreel to marketing tools.

Inside this thread, you will get free:
→ guide to creating a pro showreel in LTX-2
→ My original, royalty-free music track (+ PDF license)
→ custom transitions pack
→ curated font-pairing guide
→ Media Kit template (Figma) to land clients.

Everything you need to go from 0 to a paying client.

Save the blueprint 🧵 The Showreel:

You need it to
• show your best work
• attract partnerships
• attract new clients.

I used 2 methods:
1⃣ Text to video
2⃣ Image to video
and brought them to life with LTX-2.

Why LTX:
✅ 4k quality
✅ 50fps
✅ 20 seconds video
✅ fast gen speed

Here's how 👇Image
Oct 18, 2025 10 tweets 6 min read
Stop making content.

Start creating concepts that get you noticed by top brands.

10 years working with Warner Music, PepsiCo, and other majors taught me:
fresh ideas open every door.

I built a fresh video concept for you in LTX Studio:
5 brand logos made from… a crowd.

Steal my workflow, prompts & action plan to attract big brands with it 👇 Example 1 - Apple
Imagine pitching this to Apple for their next campaign.

Brands want fresh, scalable, authentic content that creators can deliver.

Act now:
👉 Copy my base prompt below
👉 Paste it into ChatGPT/Claude
👉 Ask: “Recreate this base prompt for [BRAND], make the whole concept and details relevant to this brand to highlight its strengths, values and tone”
👉 Generate images/videos in @LTXStudio with this new prompt
🔄 Repeat for other brands & concepts below

👉 Base prompt to copy:
Aerial nighttime photograph of a giant glowing [Apple logo (bitten apple silhouette)] formed by thousands of people on a dark city square, European architecture surroundings. Each person holding an iPhone with a flashlight creating bright white illumination. The entire logo radiating white light against a deep black background, bite area prominently lit. Directly overhead drone perspective, dramatic contrast, cinematic lighting, 4K quality.
Sep 12, 2025 10 tweets 4 min read
The 3-second problem that makes your content unseen on X 👀

I analyzed why some posts go viral while others die. It's not luck; it's a repeatable formula

Steal this entire framework + 7 hooks ready to boost your posts 👇 What's the 3-Second Problem?

The quality of your work doesn't matter if you can't win the first 3 seconds of a user's attention

On any platform - that's the only window you get ⏲️

If you lose that battle, you've already lost the war.Image
Sep 7, 2025 13 tweets 6 min read
Steal my Midjourney prompts👇🧵

I spent hours on these experiments last week:

so now just steal them and spend minutes and create your own gems 💪

Save this thread for your future use 💾 A wide football field with green turf, seen from a high elevated perspective. Multiple identical young men in casual outfits white t-shirts, light gray pants, and white sneakers are scattered across the field, each captured in different poses: walking, stretching, bending, sitting, and standing. The composition creates a surreal, almost clone-like repetition, with clear lines of the field enhancing the geometric structure. Bright daylight, sharp shadows, minimalistic and conceptual editorial photography styleImage
Aug 23, 2025 5 tweets 3 min read
Brands pay agencies $5,000+ for product shots like this

I made it:
• in 60 seconds
• for $0
• from a crappy phone photo
• using only ChatGPT

Here’s how you can do it too:
1. Snap a photo. Any product, any phone
2. Upload to ChatGPT
3. Use my prompt

Here's your prompt 👇Image Prompt:
(change text in [brackets] to your product, surrounding & colors)

Create a hyper-realistic product photo using my exact product image 9:18 (keep all text, logo, and colors 100% intact and fully readable). Make it look like a large [fire flames] has been thrown from the right side, creating a wide, powerful [energetic hot] look that wraps naturally around the product, with [fire] flowing across the front and some [fire particles] scattered around. The [fire] should feel energetic, chaotic, and natural — not too perfect or isolated. Use sharp studio lighting that highlights the [fire] motion and keeps the product text clear and crisp. The background should be a smooth gradient, [hot red to orange], with enough contrast to make the product pop. Keep the overall look premium and clean.
Aug 22, 2025 4 tweets 3 min read
Netflix said "yes" to AI 👀

The "AI is just a toy" era is over

Netflix just dropped its official guide for using Generative AI in production (link below)

It's a clear set of rules for how it will be used to create the shows you watch.

Here's what you should know 👇🧵Image Here are their 4 key insights:

1. The Creator is King:
AI is just a creative tool. In fact, it’s still people who shape and own the final story.

2. It's coming:
Netflix allows the use of GenAI in areas like concept art, VFX, and dubbing - as long as it enhances, not replaces, real creativity.

3. No stealing:
Using copyrighted material in prompts is strictly forbidden. Originality is paramount.

4. Full transparency:
You MUST disclose when and how AI was used. No hiding.Image
Aug 15, 2025 8 tweets 5 min read
I’ve got 200k+ views & recognition from Google Gemini from a single Reel 👀

In 11 days I got:
• 260 new followers
(I had ~700)
• 300+ warm leads in DMs
• 1400 saves

As a non-Insta guy, I’m surprised

Here are 3 insights I learned from this case 👇Image 1. Visual is king 🤷‍♂️

It is obvious that visually it looks really cool

This video
1️⃣ is dynamic
(from the first second)

2️⃣ builds controversy
(huge humans’ foots compared to the tiny car look weird and make you want to figure out what the hell is going on)

3️⃣ gives you recognizable aesthetics
(winter NYC is extremely well-known, so leveraging famous locations can be a good trigger for a viewer)

A few more things that can improve your videos’ performance:

👉 Catchy hook
(be careful tho, one wrong move can hurt feelings of thousands of people - when I wrote “I just created a game” in my hook on X, I got tons of hate from gaming & 3D design communities. Though, now this post almost hit 1M views - so you decide for yourself 🤷‍♂️)

👉 Sound design increases retention
(And Veo3 handles it pretty well)

👉 Offer viewers to comment
And share your prompts with them automatically

Here’s how 👇
Aug 12, 2025 5 tweets 2 min read
Why do 90% of your Midjourney loop videos are not looping?

You're likely making the same simple mistake:
❌ it's not your prompt
❌ it's not the version you're using
The secret is something else.

Here's the one mistake everyone makes - and how to avoid it 👇 The problem:

Midjourney can't loop LINEAR motion
(e.g., a car driving off-screen).

It only excels at NATURALLY REPETITIVE actions
Jul 30, 2025 12 tweets 6 min read
Now you can clone yourself - just with 1 photo (learn how 👇🧵)

❌No need to train AI on 50 photos
❌No need to spend any time on learning

Just ONE photo + prompt = you're a
• movie star
• fashion model
• FC Barcelona player

Here's how I did that (prompts below) 👇 You need just 1 minute to create your clone and start generated endless gems

1. Open the @ideogram_ai mobile or web app
2. Click "Create" -> "Character"
3. Upload your photo
4. Pick a style (or write your on prompt - steal my prompt in the posts below 👇)
5. Generate
6. Enjoy!

PS the track in the video above is made by me (no AI hehe) - listen to the full song here:
tr.ee/QADhMdHATQ

Now the sweetest part - my prompts (steal them for free):
Jul 25, 2025 14 tweets 9 min read
My Midjourney prompts + srefs thread 👇🧵

I spent hours on these experiments last week:

so now just steal them and spend minutes and create your own gems 💪

Save this thread for your future use 💾 Prompt:

[High top angle wide mid close up tracking shot, flying very fast two meters high over the grass, dominated by a real young boy (pink t-shirt, pink shorts, whit shoes, white long socks), with his back to the camera, body stretched out, gliding smoothly forward, flying in the air, casting a clear shadow on the ground. His legs and body are high above the level of the ground, his feet is not touching the ground, he is flying in the Superman pose. The background is a vast lush green valley with dense grass and dark green forest, and in the distance rolling mountains with a winding path. Large fluffy white clouds float in the vibrant blue sky. Strong dynamic motion blur creates a strong sense of speedy flight and perspective depth. Realistic image --raw]
Jul 23, 2025 12 tweets 5 min read
These AI threads brought me 400k+ views in total

No boost, no ads
• 10k+ bookmarks
• 1k+ new followers

Combined, they generated:
• 400K+ views
• 4K+ likes

Each thread is a ready-made guide that will save you hours of work

Save this mega-thread and learn with us 👇 1/10

The video that got me 150+ followers in 2 days

You'll need:
• 2-4 hours of work
• a couple of AI tools

⬇️

viral content
(72K views are insane for an account as mine, so you do the math)

Full breakdown here:
Jul 17, 2025 13 tweets 11 min read
Midjourney moody prompts thread 👇🧵

I spent hours on these experiments last week:

so now just steal them and spend minutes and create your own gems 💪

Save this thread for your future use 💾 👉 Prompt 1/11

[Pink BMW Z4M 2011 driving across soft, dense clouds, shown at a wide-angle 3/4 view, no ground visible, floating in the sky with aerial perspective, no driver inside, moody overcast atmosphere, gray and blue tones dominate the scene, subtle cinematic lighting, headlights glowing warm yellow-orange, highly realistic, dramatic composition --ar 51:92 --motion high --video 1]

Music:
open.spotify.com/track/3RyFEGqe…
Jul 8, 2025 5 tweets 2 min read
Gucci summer 👜

--sref 2334344648

Steal this Midjourney ref and show me your results - who's best in fashion photography? 👀Image
Image
@mushichstudio Don't forget to steal my previous pipeline 👇

Jul 4, 2025 8 tweets 6 min read
This video got me 150+ subs in a couple days.
Here's how I made it

It will take you 2 to 4 hours to for everything

All you need is
• an idea which has potential to pop
• couple of AI tools (free or cheap)
• any editing software

Ready?
Let's learn how to make it 👇 Step 1: Idea & Concept 🧠

1️⃣ What you like? What others like?

Brands that are "cool" in the eyes of the audience is sth that always interested me

I am also a founder of @ArtChains_ where I'm in touch with artists and... I like art

Let's cross Adidas & iconic paintings then!Image
Image
Jun 20, 2025 10 tweets 4 min read
Midjourney Video + 1 hour of work = A cartoon 💎

Here's the style reference for such an aesthetics:
--sref 2619557613

Prompts below 👇 Scene 1 prompt:

An view of the huge futuristic and dreamy building of a playful, sketchy 90's gaming center, with a crowd surrounding it and standing in line in front of its entrance,, bold graphic shapes, soft pastel background, textured crayon strokes. All the objects are edgy, fun, full of attitude, like exaggerated doodles with a retro, 90s aesthetic. --sref 2619557613Image