# Static Ads

> Generate production-ready static ad images for any brand using Claude + Nano Banana 2. End-to-end automated workflow from brand research → prompt generation → image generation via FAL API. Trigger on requests to create static ads, generate ad creatives, build ad images, or when user mentions Nano Banana, Higgsfield, FAL, or static ad generation. Also trigger when user drops a brand name + URL and asks for ad creatives.

- Skill: `sohnajonin/static-ads` (Agent Skill)
- Install (CLI): `npx skillmds@latest add sohnajonin/static-ads`
- Raw SKILL.md: https://api.skillmd.com/api/skills/sohnajonin/static-ads/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: sohnajonin (https://skillmd.com/u/sohnajonin)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/sohnajonin/static-ads

---


# Static Ad Generator (Claude Code + Nano Banana 2)

Generate 50 production-ready static ad images for any brand — from brand research to finished creatives — entirely inside VS Code.

## Overview

This skill replaces the manual Claude → Higgsfield workflow with a fully automated pipeline:

1. **Phase 1: Brand Research** → Firecrawl downloads brand images + screenshots → Claude visually inspects for accurate colors → builds Brand DNA document
2. **Phase 2: Prompt Generation** → Claude fills 50 template prompts with brand-specific details → outputs `prompts.json`
3. **Phase 3: Image Generation** → Node.js script fires prompts to Nano Banana 2 via FAL API → downloads images + builds HTML gallery

No Higgsfield needed. No copy-pasting prompts. One command, 50+ ads.

## Prerequisites

- **FAL API key** stored in `.env` file at project root (already configured)
- **Node.js 18+** (no additional packages needed — uses built-in `fetch`)
- **Product images** dropped in the brand's `product-images/` folder before running Phase 3

## Folder Structure

```
brands/{brand-name}/
├── product-images/          # Drop product PNGs/JPGs here before running
│   ├── product-front.png
│   ├── product-angle.png
│   └── ...
├── brand-images/            # Auto-downloaded from website (Step 2 + 2b)
│   ├── homepage-screenshot.png
│   ├── logo-full.png
│   ├── product-hero.png
│   ├── lifestyle-hero.jpeg
│   ├── favicon.ico
│   └── image-index.md       # Index of all downloaded images
├── brand-dna.md             # Generated by Phase 1
├── prompts.json             # Generated by Phase 2
└── outputs/                 # Generated by Phase 3
    ├── 01-headline/
    │   ├── prompt.txt
    │   ├── 1x1/
    │   │   ├── headline_1x1_v1.png
    │   │   └── headline_1x1_v2.png
    │   └── 9x16/
    │       ├── headline_9x16_v1.png
    │       └── headline_9x16_v2.png
    ├── 02-offer-promotion/
    │   └── ...
    └── gallery.html          # Visual gallery of all generated ads (grouped by ratio)
```

## How to Use

When the user invokes this skill, follow these steps:

### Step 1: Gather Inputs

Parse the arguments for:
- **Brand Name** (required)
- **Brand URL** (required)
- **Product Name** (optional — if not provided, ask the user)
- **Pricing & Offer Details** (required) — Ask the user:
  - What is the advertised price and billing structure? (e.g., "$99/mo billed annually at $1,188")
  - Does the price need a qualifier? ("starting at" / "as low as" / flat monthly)
  - Is there an upfront payment? (annual billing paid in full, setup fees, etc.)
  - Any introductory/promotional pricing vs ongoing rate?
  - Subscription terms: auto-renewal, minimum commitment, cancellation policy?
  - Geographic restrictions on pricing or availability?
  - Required legal/regulatory disclaimers for advertising? (especially health, finance, supplements)

Ask the user to drop product images into `brands/{brand-name}/product-images/` if they have them. Create the folder structure:
```
brands/{brand-name}/
└── product-images/
```

### Step 2: Download Brand Images from Website

After creating the folder structure, scrape the brand URL (and product pages) using Firecrawl with `--format images --json` to extract all image URLs from the site.

**Process:**
1. Run `firecrawl scrape "<brand_url>" --format images --json` on the homepage and key product/service pages
2. Filter the returned image URLs — keep only brand-relevant images:
   - Product photos, packaging, hero shots
   - Lifestyle/campaign photography
   - Logo files
   - Skip: generic icons, widget images, payment card logos, CDN placeholder SVGs, avatars, third-party badges
3. Download each relevant image using `curl -sL -o <filename> "<url>"` into `brands/{brand-name}/brand-images/`
4. Name files descriptively (e.g., `trt-single-vial.png`, `lifestyle-hero.jpeg`, `logo-dark.png`)

**These downloaded images serve as:**
- Direct reference images to attach in Nano Banana 2 alongside ad prompts
- Visual inspiration for prompt generation
- Brand color/style validation for the Brand DNA document

Save an `image-index.md` file in the `brand-images/` folder listing each downloaded image with a short description of what it shows.

### Step 2b: Visual Brand Inspection (Screenshots)

**This step prevents color/branding mistakes.** Text-only web scraping cannot reliably capture visual brand identity — colors described in CSS may not match the actual rendered brand palette, and logos/product details are invisible to text scrapers. Screenshots solve this.

**Process:**
1. Take a full-page screenshot of the homepage:
   ```
   firecrawl scrape "<brand_url>" --full-page-screenshot --json
   ```
2. Parse the JSON output to extract the screenshot URL (it's a Google Cloud Storage PNG URL in the `screenshot` field)
3. Download the screenshot:
   ```
   curl -sL -o "brands/{brand-name}/brand-images/homepage-screenshot.png" "<screenshot_url>"
   ```
4. Take additional screenshots of key pages (product page, about page) if they have distinct visual treatments:
   ```
   firecrawl scrape "<product_page_url>" --screenshot --json
   ```
5. Download the favicon separately:
   ```
   curl -sL -o "brands/{brand-name}/brand-images/favicon.ico" "<brand_url>/favicon.ico"
   ```

**After downloading screenshots, visually analyze them using Claude's multimodal capability:**
- Read each screenshot image file to visually inspect it
- Extract the **actual rendered brand colors** — primary, secondary, accent, background, CTA button colors
- Identify **logo details** — icon shape, wordmark style, color usage on light vs dark backgrounds
- Note **typography** — headline weight/style, body text, any distinctive treatments
- Observe **photography style** — lighting, color grading, subject matter, mood
- Check **overall aesthetic** — clean/minimal vs bold/dense, color temperature, spacing

**CRITICAL:** The visual analysis from screenshots is the **primary source of truth** for brand colors and visual identity in the Brand DNA document. Text-scraped CSS/color references are secondary. If there's a conflict between what text scraping suggests and what the screenshots actually show, **always trust the screenshots.**

Save screenshots to `brand-images/` and add them to the `image-index.md` index file.

### Step 3: Brand DNA Research (Phase 1)

Use web search/scraping tools extensively to gather real brand data. Execute the following research process:

```
Role: Act as a Senior Brand Strategist conducting a full reverse-engineering of the target brand's visual and verbal identity.

Objective: Create a comprehensive Brand DNA document that will be used to write highly specific AI image generation prompts. Every detail matters because the output will be fed into an image model that needs exact specifications.

PHASE 1: EXTERNAL RESEARCH

Use web search to find the source of truth for this brand:

1. Design credits: Search for "who designed [Brand] branding", "[Brand] design agency case study", "[Brand] rebrand"
2. Public brand assets: Search for "[Brand] brand guidelines pdf", "[Brand] press kit", "[Brand] media kit", "[Brand] style guide"
3. Typography: Search for "[Brand] font", "[Brand] typeface", "what font does [Brand] use"
4. Colors: Search for "[Brand] brand colors", "[Brand] hex codes", "[Brand] color palette"
5. Packaging: Search for "[Brand] packaging design", "[Brand] unboxing", "[Brand] product photography"
6. Advertising: Search "[Brand]" in Meta Ad Library (facebook.com/ads/library) for current ad creative styles
7. Press and positioning: Search for "[Brand] brand story", "[Brand] founding story", "[Brand] mission"

PHASE 2: VISUAL ANALYSIS (from screenshots captured in Step 2b)

Read the homepage screenshot and any additional page screenshots captured earlier. Use these as the PRIMARY source of truth for all visual brand attributes:

1. Brand Colors (from screenshots): Identify the ACTUAL rendered primary color, secondary color, accent color, background colors, and CTA button color. Extract approximate hex codes from what you see. DO NOT rely solely on text-scraped CSS — verify against the visual screenshot.
2. Logo Details (from screenshots + downloaded logo files): Describe the exact logo — icon shape, colors used, wordmark font style, any distinctive elements (e.g., stylized letters, icon marks).
3. Photography Style: Describe lighting, color grading, composition, and subject matter visible in hero images and product photography on the site.
4. Typography on site: Headline weight, body weight, letter-spacing, distinctive treatments visible in the screenshots.
5. Color application: How are colors actually used? Primary vs accent. Background colors. CTA button color and shape.
6. Layout density: Airy or dense? Grid-based or organic?
7. Packaging details: Physical appearance (materials, colors, shape, label placement, textures, translucency, matte vs gloss) from product images.

PHASE 2b: ON-SITE TEXT ANALYSIS

Visit the Target URL via text scraping and analyze:

1. Voice and Tone: Read hero copy, About page, and product descriptions. Give me 5 distinct adjectives.
2. Cross-reference any color/font information found in CSS or text with what you observed in the screenshots. Screenshots win if there's a conflict.

PHASE 3: COMPETITIVE CONTEXT

Search for 2-3 direct competitors and note visual differentiation.

PHASE 4: OUTPUT

Combine into this format:

BRAND DNA DOCUMENT
==================

BRAND OVERVIEW
Name / Tagline / Design Agency / Voice Adjectives [5] / Positioning / Competitive Differentiation

VISUAL SYSTEM
Primary Font / Secondary Font / Primary Color [hex] / Secondary Color [hex] / Accent Color [hex] / Background Colors / CTA Color and Style

PHOTOGRAPHY DIRECTION
Lighting / Color Grading / Composition / Subject Matter / Props and Surfaces / Mood

PRODUCT DETAILS
Physical Description / Label-Logo Placement / Distinctive Features / Packaging System

AD CREATIVE STYLE
Typical formats / Text overlay style / Photo vs illustration / UGC usage / Offer presentation

PRICING & OFFER STRUCTURE
Advertised price / Billing cadence (monthly vs annual vs one-time) / Actual upfront cost / Price qualifier language ("starting at" / "as low as" / flat) / Introductory offers vs ongoing rate / Subscription terms / Required ad disclaimers / Why this structure matters for ad compliance

IMAGE GENERATION PROMPT MODIFIER
Write a single 50-75 word paragraph to prepend to any image prompt to match this brand's visual identity. Include exact colors, font descriptions, photography direction, and mood.
```

Save output as: `brands/{brand-name}/brand-dna.md`

### Step 3b: Confirm Reference Images

**STOP and ask the user before proceeding to prompt generation.** The user must confirm which images in `product-images/` should be used as reference images for the generation script.

**Process:**
1. List all files currently in `brands/{brand-name}/product-images/` with their filenames
2. Ask the user:
   - Which of these images should be sent as reference images to the AI model?
   - Should all images be used for every prompt, or should specific images be assigned to specific templates?
   - Are there any additional images they want to add before proceeding?
3. Show the full file paths so the user can verify (e.g., `brands/{brand-name}/product-images/Tirzepatide-Vial.png`)
4. Based on the user's answer, note which images map to which prompts — this will populate the `reference_images` array in each prompt entry in `prompts.json`

**Reference image rules for prompts.json:**
- If the user wants ALL images sent with every prompt: leave `reference_images` as an empty array `[]` (the script sends all product images by default when the array is empty)
- If the user wants SPECIFIC images per prompt: populate `reference_images` with an array of filenames (just the filename, not full path) — e.g., `"reference_images": ["Tirzepatide-Vial.png", "product-front.png"]`
- The filenames must match exactly what's in `product-images/`

**Do NOT proceed to Step 4 until the user confirms their reference image selection.**

### Step 4: Generate Prompts (Phase 2)

After the Brand DNA is complete and reference images are confirmed, generate brand-specific prompts from the 50 templates below.

For each template:
1. Replace all [BRACKETED PLACEHOLDERS] with brand-specific details from the Brand DNA
2. Prepend the Image Generation Prompt Modifier from the Brand DNA
3. **Do NOT set an `aspect_ratio`** — the script automatically generates both 1:1 and 9:16 for every prompt
4. Include the product name and any specific product details

**Output as a JSON file** with this structure and save to `brands/{brand-name}/prompts.json`:

```json
{
  "brand": "Brand Name",
  "product": "Specific Product Name",
  "generated_at": "ISO timestamp",
  "prompts": [
    {
      "template_number": 1,
      "template_name": "headline",
      "prompt": "Full completed prompt text ready for Nano Banana 2...",
      "notes": "Any generation notes"
    }
  ]
}
```

**Important notes for the generated prompts:**
- Do NOT include aspect ratio text in the prompt itself (the script generates both 1:1 and 9:16 automatically)
- Make sure all copy in prompts is realistic and brand-appropriate
- Use actual brand colors (hex codes) from the Brand DNA
- Reference specific product details from the research
- Each prompt should be self-contained and ready to fire
- Escape any double quotes in the prompt text for valid JSON
- **Pricing language compliance:** Always use the exact price qualifier from the Brand DNA PRICING STRUCTURE section (e.g., "starting at $99/mo" NOT "$99/mo"). If pricing is conditional on billing term, commitment, or tier, the qualifier is mandatory. Include billing context ("*billed annually", "(annual plan)") where ad copy has room. Presenting conditional pricing as flat pricing creates "bait and switch" risk on paid ad platforms.
- **Ad compliance checklist:**
  - Testimonials/reviews: Include "Individual results may vary" disclaimer where space allows
  - Before/after claims: Especially regulated in health/wellness — add small disclaimer text
  - Competitor comparisons: Ensure pricing claims are defensible and current
  - Geographic availability: Note if not available in all states/regions
  - Medical/health products: Include appropriate regulatory disclaimers
  - Promotional offers: Clarify terms (limited time, new customers only, etc.)

### Step 5: Generate Images (Phase 3)

Copy the generation script from `skills/references/generate_ads.mjs` to the brand folder, then run it.

```bash
# Copy the script
cp skills/references/generate_ads.mjs brands/{brand-name}/generate_ads.mjs

# Run for ALL templates (generates ~160 images at 4/prompt)
node brands/{brand-name}/generate_ads.mjs --brand-dir brands/{brand-name}

# Or run for SPECIFIC templates (cheaper for testing)
node brands/{brand-name}/generate_ads.mjs --brand-dir brands/{brand-name} --templates 1,7,13,15

# Options:
#   --templates 1,7,13    Only generate specific template numbers
#   --num-images 2        Generate 2 images per prompt (default: 4)
#   --resolution 1K       Use 1K resolution (default: 2K, options: 0.5K/1K/2K/4K)
```

The script will:
1. Read `prompts.json` from the brand folder
2. For each prompt, generate **both 1:1 and 9:16 aspect ratios** via `fal-ai/nano-banana-2` (text-to-image)
3. Run all jobs in parallel (configurable with `--max-concurrent`, default 10)
4. Download generated images into `outputs/{template-number}-{template-name}/1x1/` and `9x16/`
5. Save the prompt text alongside each template
6. Generate `outputs/gallery.html` showing all generated ads grouped by ratio

### Step 6: Present Results

After generation completes, tell the user:
- How many images were generated across how many templates
- Where the output folder is located
- That they can open `gallery.html` in a browser to review all ads
- Any templates that failed (and offer to retry them)

Ask if they want to:
- Generate additional templates
- Regenerate specific templates with tweaked prompts
- Run for a different product using the same Brand DNA

## User Interaction Flows

### Quick Start: "generate ads for [brand]"
1. Ask for: brand name, URL, specific product name
2. Confirm they have product images ready (or ask them to drop images in the folder)
3. Run Steps 1-6: Gather inputs → Download brand images → Screenshot + visual inspect → Brand DNA → Prompts → Image generation
4. Present the output gallery

### Selective Generation: "just generate templates 1, 7, 13"
1. Confirm brand folder and prompts.json exist
2. Run Phase 3 with `--templates` filter

### New Product, Same Brand: "now do it for [different SKU]"
1. Skip Phase 1 (Brand DNA + brand-images already exist)
2. Re-run Phase 2 with new product details
3. Run Phase 3

### Prompt-Only Mode: "just give me the prompts, I'll run them in Higgsfield"
1. Run Phase 1 + Phase 2 only
2. Present prompts in markdown format (not JSON)
3. User copies prompts manually to Higgsfield/Nano Banana 2 playground

## FAL API Technical Details

- **Model**: `fal-ai/nano-banana-2` (text-to-image)
  - Input: prompt, aspect_ratio, num_images, output_format, resolution, safety_tolerance
- **Aspect ratios**: auto, 21:9, 16:9, 3:2, 4:3, 5:4, 1:1, 4:5, 3:4, 2:3, 9:16
- **Resolutions**: 0.5K, 1K, 2K, 4K (default: 2K for production quality)
- **API authentication**: `Authorization: Key {FAL_KEY}` header
- **Queue API**: POST to `https://queue.fal.run/fal-ai/nano-banana-2`, poll status, fetch result
- **Dual aspect ratios**: Every prompt generates both 1:1 and 9:16 automatically. Output goes into `1x1/` and `9x16/` subfolders.
- **Parallel generation**: `--max-concurrent N` controls simultaneous API calls (default 10)
- **Cost**: ~$0.08/image at 1K, ~$0.12/image at 2K, ~$0.16/image at 4K
  - Full 50-template run at 4 images each at 2K (x2 ratios) = ~$48.00

## Template Prompts Library

### 1. Headline
Tests text rendering. If this comes back clean, the model handles copy.

Template: Use the attached images as brand reference. Match the exact product colors, typography style, and brand tone precisely. Create: a static ad with a [BACKGROUND] background. Top third: large bold sans-serif headline reading "[YOUR HEADLINE, under 10 words]". Below in smaller text: "[YOUR SUBHEAD, one sentence]". Bottom half: [YOUR PRODUCT] on the surface with [DETAILS]. Shot at 50mm f/2.8 from slightly above. [BRAND] logo bottom right. Clean, authoritative.

### 2. Offer/Promotion
The money-maker. Test your core offer.

Template: Use the attached images as brand reference. Match exact brand colors and typography style. Create: a promotional ad with a split background. Top 60% is [PRIMARY BRAND COLOR] and bottom 40% is [CONTRAST COLOR like warm cream]. [YOUR PRODUCT] sits centered where colors meet, soft studio lighting. Upper area: large [CONTRAST TEXT] sans-serif reading "[YOUR OFFER like YOUR FIRST MONTH FREE]". Below: "[OFFER DETAILS]". Lower section: small [BRAND COLOR] text with [VALUE ADDS]. [BRAND] logo bottom right.

### 3. Testimonials
Real environments + text overlays. Tests composition depth.

Template: Use the attached images as brand reference. Create: a testimonial ad set in [SETTING like bright bathroom / kitchen] with warm natural light. [YOUR PRODUCT] on [SURFACE], slightly out of focus. Overlaid: large bold white sans-serif "[SHORT HEADLINE]". Below: "[FULL QUOTE 2-3 sentences]. [NAME], [CREDENTIAL]." Five filled [BRAND COLOR] stars. [BRAND] logo bottom right in white. Shot on 35mm f/2.0.

### 4. Features/Benefits Point-Out
Educational diagram-style layout.

Template: Use the attached images as brand reference. Create: an educational diagram-style ad on white background. Top: bold [BRAND COLOR] text "[HEADER like What Makes [PRODUCT] Different]". Below: [YOUR PRODUCT] centered, even studio lighting. Four callout boxes with connecting lines: "[BENEFIT 1-4]". Each has a small [BRAND COLOR] circle. "[WEBSITE]" bottom center. [BRAND] logo bottom right. Scientific diagram redesigned by a luxury agency.

### 5. Bullet-Points
Split composition. Product left, benefits right.

Template: Use the attached images as brand reference. Create: a benefit-list ad, split composition on [BACKGROUND] background. Left 40%: [YOUR PRODUCT] on [SURFACE], shot at 85mm f/2.8. Right 60%: vertical stack of five lines with filled [BRAND COLOR] circles: "[BENEFIT 1-5]". Clean sans-serif, generous spacing. [BRAND] logo bottom right.

### 6. Social Proof
Member count + review card + press logos. The trust stack.

Template: Use the attached images as brand reference. Create: a social proof ad on [BACKGROUND like warm cream]. Top: "[HEADLINE like Join 1,000,000+ Members]" in bold [BRAND COLOR]. Five filled stars with "Rated [X] out of 5". Center: [YOUR PRODUCT] at 50mm f/4. Below: frosted white card with five-star rating, "[REVIEW TITLE]", "[2-3 SENTENCE REVIEW]", "[ATTRIBUTION]" in italic. Below card: "As Featured In" with five grayscale logos. [BRAND] logo bottom right.

### 7. Us vs Them
Side-by-side comparison. Photography quality gap IS the argument.

Template: Use the attached images as brand reference. Create: a side-by-side divided vertically. Left: muted gray-blue background. Right: [PRIMARY BRAND COLOR]. Center top: white circle with "VS". Left header: "[COMPETITOR CATEGORY]" + generic competitor product + list with X marks: "[WEAKNESS 1-5]". Right header: "[YOUR BRAND]" + [YOUR PRODUCT] + list with checkmarks: "[STRENGTH 1-5]". [BRAND] logo bottom right.

### 8. Before & After (UGC Native)
Mirror selfie transformation. Must look like a real person posted it.

Template: Use the attached images as brand reference for product color ONLY. This should look like a real person's post. Create: TikTok before-and-after. LEFT: grainy iPhone mirror selfie, [PERSON] in dimly lit bathroom, [BEFORE STATE], harsh lighting. White handwritten text: "[BEFORE DATE]". RIGHT: same person, same bathroom, bright natural light, [AFTER STATE], [PRODUCT] visible on counter. White text: "[AFTER DATE]". Top center: "[TIMEFRAME] on [BRAND]" with emoji. Should look stitched in CapCut.

### 9. Negative Marketing (Bait & Switch)
Fake bad review that's actually a rave. Scroll-stopper.

Template: Use the attached images as brand reference. Create: Background is close-up of [PRODUCT], slightly blurred. Center: white rounded-rectangle review card (Amazon-style). Gray user icon, "[NAME]", one gold star + four gray, orange "Verified Purchase" badge, bold text: "[BAIT that sounds negative but is positive]". Bottom: bold white sans-serif "[PUNCHLINE like THE REVIEWS ARE IN.]". [BRAND] logo bottom right.

### 10. Press/Editorial
Authority play. Vogue back-page energy.

Template: Use the attached images as brand reference. Create: a press ad on off-white linen background. Top: "As Featured In" in small [BRAND COLOR] uppercase wide-tracked text. Below: five grayscale publication logos. Center: italic serif pull-quote in [BRAND COLOR]: "[PRESS QUOTE]" with attribution. Lower third: [PRODUCT] at 85mm f/2.8, soft side light. [BRAND] logo bottom left. Generous white space. Full-page Vogue energy.

### 11. Pull-Quote Review Card
Emotional quote headline over a truncated review card on a color block. The "...Read more" creates an open loop.

Template: Use the attached images as brand reference. Match the exact product colors and brand tone precisely. Create: a review-driven ad with a solid [BRAND COLOR with hex] color block background filling the entire image. Top half: large bold italic serif text in white with curly quotation marks reading "[PULL-QUOTE — the most emotional 4-8 word phrase from the review]". Directly below the quote: five large filled gold/yellow star icons in a horizontal row. Bottom left, overlapping the color background: a white rounded-corner review card with subtle shadow, containing: a small gray circular default avatar icon, beside it "[FIRST NAME + LAST INITIAL]" in bold dark sans-serif with a small flag emoji, below the name a blue checkmark icon with "Verified Reviewer" in small blue text. Below the reviewer info: the review body text in medium-weight dark sans-serif, 4-6 lines of authentic-sounding customer voice that trails off mid-sentence, ending with "...Read more" in bold [BRAND COLOR] text. Below the review text: "Was this review helpful?" in small gray text with a thumbs-up icon and "[HELPFULNESS COUNT]" beside it. Bottom right, overlapping both the card and the color background: [YOUR PRODUCT] angled slightly toward the viewer, sitting on the color block surface with a subtle shadow beneath. No brand logo needed if the product packaging already shows it.

### 12. Lifestyle Action + Product Colorway Array
Action hero shot sells the use case. Fanned product lineup sells the range.

Template: Use the attached images as brand reference. Match the exact product design, colors, and visual tone precisely. Create: a static ad with a [LIFESTYLE PHOTO DESCRIPTION] occupying the left two-thirds of the frame, shot outdoors in [SETTING], bright natural daylight. [BRAND] logo top center in bold. Below logo: large bold sans-serif quote text reading "[ENDORSEMENT HEADLINE]" in [TEXT COLOR]. Bottom right foreground: three [PRODUCT VARIANTS] fanned in an overlapping arrangement showing [COLOR 1], [COLOR 2], and [COLOR 3]. Products are crisp and studio-lit against the lifestyle background. Shot on 50mm f/2.0, lifestyle background slightly softer than foreground product.

### 13. Stat Surround / Callout Radial (Product Hero)
Product is the sun. Stats are the planets. Arrows make it scannable in under 2 seconds.

Template: Use the attached images as brand reference. Match the exact product design, colors, and typography style precisely. Create: a static ad on a white-to-[LIGHT GRADIENT COLOR] gradient background, gradient fading from top to bottom. Top: large bold [TEXT COLOR] sans-serif headline reading "[HEADLINE]" Center: [YOUR PRODUCT] on white background, soft studio lighting. Floating near the product: a small circular badge reading "[PRICE POINT]" in [BADGE COLOR]. Flanking the product on both sides: four stat callouts with curved arrows pointing toward the product. Left side top: "[STAT 1]" in oversized bold text with "[LABEL]" below. Left side bottom: "[STAT 2]" with "[LABEL]". Right side top: "[STAT 3]" with "[LABEL]". Right side bottom: "[STAT 4]" with "[LABEL]" and five filled gold stars beneath. Arrows are simple hand-drawn-style curved lines. Bottom foreground: [FLAVOR PROPS] adding appetite appeal. No brand logo. Clean, informational, appetizing.

### 14. Bundle Showcase + Benefit Bar
Sells the system, not the SKU. The open box is the hero.

Template: Use the attached images as brand reference. Match the exact product design, colors, and typography style precisely. Create: a static ad on a [BACKGROUND like soft gradient] background. Top: oversized bold white all-caps sans-serif headline reading "[HEADLINE]". Below headline: a horizontal [ACCENT COLOR] banner bar divided into [NUMBER] equal segments separated by thin vertical lines, each containing a two-word benefit label in white text: "[BENEFIT 1-5]". Center-to-bottom: an open [PACKAGING] photographed at a slight overhead angle showing [NUMBER] [PRODUCTS] nestled inside. In the lower foreground: a [LIFESTYLE PROP] entering the frame from bottom. [BRAND] logo bottom left corner. Bright studio lighting, saturated color, energetic.

### 15. Social Comment Screenshot + Product
Screenshotted comment = instant credibility. Looks organic.

Template: Use the attached images as brand reference. Match the exact product design and colors precisely. Create: a static ad on a clean white background. Top: oversized bold black sans-serif headline reading "[HOOK HEADLINE]" with emoji at the end. Center: a social media comment card with light gray rounded-rectangle background containing: a small circular profile avatar (top left), bold name "[REVIEWER NAME]", and a multi-sentence review in regular-weight sans-serif: "[FULL REVIEW TEXT, 3-4 sentences, conversational and emotional]". Small gray timestamp below the comment. Bottom center: [YOUR PRODUCT] photographed at a slight angle on white, soft studio lighting. No brand logo. No stars. Should look like someone screenshotted a real comment and dropped the product photo below it.

### 16. Curiosity Gap / Hook Quote Testimonial
Provocative headline forces the double-take. The reveal reframes it.

Template: Use the attached images as brand reference. Match the exact product design, colors, and typography style precisely. Create: a static ad on a clean white background. Top center: large [ACCENT COLOR] opening quotation marks. Below: mixed-weight headline in black — the first line in italic serif reading "[SETUP LINE]", the next two lines in enormous heavy-weight bold all-caps sans-serif reading "[BAIT PHRASE]", followed by a smaller sentence-case line reading "[REVEAL]". Closing quotation marks and "[ATTRIBUTION]" in regular weight. Left side bottom third: [YOUR PRODUCT] at a slight angle with [PRODUCT DETAILS]. To the left of the product: a [TRUST BADGE]. Right side bottom third: [NUMBER] filled [ACCENT COLOR] stars and bold text reading "[REVIEW COUNT] 5-Star Reviews". Bottom edge: small disclaimer text.

### 17. Verified Review Card
Mimics real review platform UI. The verified badge and helpfulness count do the trust-building work.

Template: Use the attached images as brand reference. Match the exact product design, colors, and typography style precisely. Create: a static ad on a solid [BRAND COLOR] background. Top: large bold white serif pull-quote reading "[HEADLINE QUOTE]" in quotation marks. Below the quote: five filled gold stars, large. Center-left: a white rounded-rectangle review card with subtle shadow containing: gray circular avatar icon, bold name "[REVIEWER NAME]" with flag emoji, blue checkmark and "Verified Reviewer" in [BRAND COLOR] text, then 3-4 sentences of review body text. At the bottom of the card: a blue "...Read more" link and "Was this review helpful?" with thumbs-up icon. Right side, overlapping the card edge: [YOUR PRODUCT] photographed at a slight angle, soft studio lighting. No brand logo. The review UI is the trust mechanic.

### 18. Stat Surround / Callout Radial (Lifestyle Flatlay)
Same stat-radial format but lifestyle flatlay background sells the experience.

Template: Use the attached images as brand reference. Match the exact product design, colors, and typography style precisely. Create: a static ad on a white background with a lifestyle flatlay arrangement. Top: bold [ACCENT COLOR] filled banner bar spanning full width, white all-caps sans-serif reading "[HEADLINE]". Center: a [PERSON DETAIL] holding [YOUR PRODUCT] in mid-frame. Scattered around the edges: [FLATLAY PROPS] arranged organically. Four stat callouts with curved [ACCENT COLOR] arrows pointing toward the held product: "[STAT 1-4]" with labels. Stats in bold black, labels in all-caps regular weight. Bright, flat studio lighting. Energetic, appetizing, information-dense but scannable.

### 19. Highlighted / Annotated Testimonial
The highlighter pen does the work. Key phrases visually emphasized.

Template: Use the attached images as brand reference. Match the exact product design, colors, and typography style precisely. Create: a static ad on a clean white background. Top left: circular customer headshot photo of [PERSON DESCRIPTION]. To the right: bold name "[REVIEWER NAME]" with a verified icon. Below: a long-form customer quote in large regular-weight black sans-serif. Key phrases highlighted with [HIGHLIGHT COLOR] rectangular background fills behind the text: "[HIGHLIGHTED PHRASE 1]", "[HIGHLIGHTED PHRASE 2]". Bottom right: [YOUR PRODUCT] at a slight angle, partially cropped at the bottom edge. To the left of the product: a circular [TRUST BADGE] seal. [BRAND] logo bottom left corner, small.

### 20. Advertorial / Editorial Content Card
Looks like a news post, not an ad. Designed to blend into feed.

Template: Use the attached images as brand reference for tone ONLY. Do NOT use polished ad layouts. This should look like organic editorial content. Create: a full-bleed moody portrait photo of [PERSON DESCRIPTION], warm amber-toned lighting, shot on 50mm f/1.8, shallow depth of field, cinematic color grade. Lower 45%: a prominent white rounded-rectangle pill label reading "[CATEGORY TAG]" in centered uppercase sans-serif. Below: very large, dominant, bold all-caps condensed sans-serif headline in white text with key words in [HIGHLIGHT COLOR]: "[HEADLINE]". The headline should be oversized. Bottom center: "[@HANDLE]" in small white text. No product shot, no CTA button, no stars. Should read like a culture account post.

### 21. Bold Statement / Reaction Headline
Pure brand energy. No stats. No reviews. Just a provocative line, a hot gradient, and the product.

Template: Use the attached images as brand reference. Match the exact product design, colors, and visual tone precisely. Create: a static ad on a vibrant [GRADIENT] gradient background, flowing diagonally from upper left to lower right. Upper left: oversized playful [FONT STYLE] white headline reading "[BOLD STATEMENT]" — text should feel loose, fun, and expressive. Right side: [PERSON DETAIL] grabbing from [YOUR PRODUCT]. Product sits center-right, slightly below midline. Bottom left: [BRAND] logo in [LOGO COLOR] with "[PRODUCT DESCRIPTOR]" in smaller text below. No stats, no reviews, no badges. The gradient and the headline do all the work.

### 22. Flavor Story / "Tastes Like"
Template: Use the attached images as brand reference. Match the exact product design, colors, and packaging precisely. Create: a flavor-visualization ad. Full background is a photorealistic close-up food scene of [INDULGENT FOOD]. Shot at 50mm f/2.8, shallow depth of field, warm bakery lighting. Top third: large bold white sans-serif headline reading "[HEADLINE]" with one key word in bold italic for emphasis. [YOUR PRODUCT] placed bottom-right, angled 15 degrees as if casually laid into the scene. Bottom: semi-transparent white bar with three stat columns: "[STAT 1]" | "[STAT 2]" | "[STAT 3]". Very bottom edge: smaller bold sans-serif "[CLEAN LABEL CLAIM]". Food is the hero — product is the payoff.

### 23. Long-Form Manifesto / Letter Ad
Template: Use the attached images as brand reference. Match exact brand typography style and tone. Create: a copy-dominant manifesto ad on a clean white background. No background imagery — text is the entire creative. Top: oversized bold black serif or sans-serif headline reading "[PROVOCATIVE HEADLINE]" spanning the top 15%. Below: left-aligned body copy in smaller regular-weight black text, structured as short punchy sentences and line breaks (NOT paragraphs). Approximately 12-18 lines of copy. Bottom 20%: [YOUR PRODUCT] centered or slightly right, product-only on white, clean studio shot angle. No icons, no badges, no CTA button. The writing IS the ad.

### 24. Product + Comment Callout (Faux Social Proof)
Template: Use the attached images as brand reference. Match the exact product design and packaging precisely. Create: a social proof ad. Top 55%: [YOUR PRODUCT] centered on a clean white background, studio-lit, shot at 85mm f/2.8 with soft shadow. A few [LOOSE UNITS] spilling casually at the base. Bottom 45%: a realistic Facebook-style comment card. Left: small circular profile photo of [PERSON DESCRIPTION]. Bold name "[FIRST NAME + LAST INITIAL]" above the comment. Comment text: "[TESTIMONIAL 2-3 sentences]". Below: "[TIMEFRAME]" - Like - Reply in gray. Bottom right: Facebook-style reaction emojis with "[COUNT]". Should look like an organic screenshot.

### 25. Us vs. Them Color Split
Template: Use the attached images as brand reference. Match the exact product design and colors precisely. Create: a side-by-side comparison ad divided vertically. Left half: [PRIMARY BRAND COLOR] background. [YOUR PRODUCT] with [ACTION DETAIL]. Brand logo in bold white upper-left. Below: vertical stack of 4 benefits with green checkmarks: "[STRENGTH 1-4]". Right half: [CONTRAST COLOR] background. Generic unbranded competitor product. Header: "[COMPETITOR CATEGORY]". Below: vertical stack of 4 weaknesses with red X marks: "[WEAKNESS 1-4]". Center divider: a comic-style "VS" burst graphic.

### 26. Stat Callout (Data-Driven Lifestyle)
Template: Use the attached images as brand reference. Match exact brand colors and typography. Create: a statistic-led ad. Top 50%: lifestyle product photography — [SCENE]. Product packaging visible. Middle: brand logo centered with thin horizontal rules. Bottom 50%: dark gradient overlay. Large bold uppercase sans-serif text: "[STAT-DRIVEN HEADLINE with specific percentages]." Key result phrases highlighted in [ACCENT COLOR].

### 27. Benefit Checklist Showcase (Split Product + Info)
Template: Use the attached images as brand reference. Match the exact product design and brand colors precisely. Create: an information-dense benefit ad, split composition. Left 45%: overhead product shot — [PRODUCT DISPLAY DESCRIPTION]. Right 55%: white background. Top: [STAR RATING] with "[REVIEW COUNT]". Brand logo. Below: [BRAND COLOR] headline: "[HEADLINE]". Then 3 checkmark benefit rows. Bottom right: large rounded [ACCENT COLOR] CTA button reading "[CTA]".

### 28. Feature Arrow Callout / Product Annotation
Template: Use the attached images as brand reference. Match exact brand colors and typography style. Create: a product annotation ad on a [BACKGROUND] background. Top: italic serif headline "[BENEFIT STATEMENT]". Below in massive bold sans-serif: "[VALUE PROP]". Center: [PERSON'S HAND] holding [YOUR PRODUCT]. Four curved arrows pointing from the product outward to four benefit callout labels. Arrows should feel hand-drawn or editorial. Bottom: full-width [CONTRAST COLOR] banner with [PROMO TEXT].

### 29. UGC + Viral Post Overlay
Template: Use the attached images as brand reference for product color ONLY. Do NOT use ad layouts or polish. Create: a casual selfie of [PERSON] doing something mundane. iPhone front camera, slightly grainy. Overlaid in the center: a realistic screenshot of a [PLATFORM] post. Post title in bold: "[PROVOCATIVE OPINION HEADLINE]". Post body: "[2-3 sentences]". Should feel like the person is reacting to it. No product visible. No brand logo. No CTA. The hook is the opinion.

### 30. Hero Statement + Icon Benefit Bar
Template: Use the attached images as brand reference. Match exact brand colors, packaging, and typography. Create: a bold statement ad. Top 15%: white banner with massive bold [BRAND COLOR] uppercase sans-serif headline: "[2-3 WORD POWER STATEMENT]" with a period. Middle 55%: lifestyle product photo — [SCENE]. Product label clearly visible. Bottom 25%: [SOFT BRAND COLOR] background. Three evenly spaced icon-and-text benefit columns. Very bottom edge: scrolling ticker bar with "[SOCIAL PROOF]".

### 31. Comparison Grid / Table
Template: Use the attached images as brand reference. Match the exact product packaging. Create: a structured comparison grid ad on white background. Top row: [YOUR PRODUCT] vs [COMPETITOR PRODUCT]. Below: three horizontal rows each comparing one attribute. All text in bold black serif or heavy sans-serif, centered. No icons, no colors, no checkmarks — the copy contrast does the work. Should feel like a meme-format comparison.

### 32. UGC Story Callout / Text Bubble Explainer
Template: Use the attached images as brand reference for product color and packaging ONLY. Must look like a real Instagram Story. Create: a casual iPhone photo of [PERSON DESCRIPTION] holding [YOUR PRODUCT] over [SURFACE/SETTING]. Natural overhead daylight. Scattered across the frame: 5 text bubbles using Instagram Story's built-in highlighted text tool with varied highlight colors. Should feel casual and hand-placed, not designed.

### 33. Faux Press / News Articles Screenshot
Template: Use the attached images a

…(truncated)