Creative Validator Docs

Upload one image, validate it against 50+ advertising platform specs, and generate correctly-sized derivatives for every channel with AI-powered generative fill. Convert still images to video with Veo 3.1. Runs in the browser — no install, no signup for validation.

Free to validate Validation is instant and requires no account. Resize uses Cloudinary (free tier available). Animate uses Google AI Studio (free Veo 3.1 access).

The Creative Asset Validator solves a specific pain point: you have one image and need it sized correctly for Google Ads, Meta, TikTok, DV360, The Trade Desk, CTV, LinkedIn, Pinterest, Snapchat, and dozens more. Manually checking specs and resizing in Photoshop takes hours. This tool does it in seconds.

What makes it different from a simple crop-and-resize tool:

AI generative fill — extends images rather than cropping them. A 1:1 square image becomes a 728×90 leaderboard by generating new background content, not by cutting the original.
Smart strategy selection — automatically picks c_fill,g_auto for close aspect ratios (minimal change) and c_pad,b_gen_fill for dramatically different ratios (generative extension).
163 channel checks — every platform, every placement, every size. Aspect ratio, pixel dimensions, file size, duration for video.
Image-to-video — convert any still image to a 4–10 second animated video ad using Google Veo 3.1.

Getting Started

Three steps from upload to production-ready assets:

Upload an Image

Drag and drop or click to upload any image (JPG, PNG, WebP, GIF) or video (MP4, MOV). The tool reads dimensions, aspect ratio, and file size instantly in the browser.

See Validation Results

Every supported channel is checked and color-coded: green = compatible as-is, amber = close but needs resize, red = incompatible aspect ratio. No account needed — validation runs entirely client-side.

AI Resize All

Click "Fix All" or select individual channels. The tool sends your image to Cloudinary with the correct transformation parameters and returns production-ready derivatives. Requires a Cloudinary account (free tier works).

BYOK — Bring Your Own Keys Cloudinary credentials (cloud name, API key) are entered in the app and stored in your browser's sessionStorage. They are never sent to any server other than Cloudinary's API. Closing the tab destroys them.

How Validation Works

When you upload an image, the validator checks it against every supported channel's requirements:

Aspect Ratio Matching

Each channel has an expected aspect ratio. The validator computes your image's ratio and compares it with a tolerance of ±0.05. For example, a 1080×1080 image (ratio 1.0) matches any channel requiring a ratio between 0.95 and 1.05.

Exact Size Matching (Display)

For display ad sizes (GDN, DV360, TTD), exact pixel dimensions are required. A 300×250 Medium Rectangle must be exactly 300×250 — aspect ratio matching alone isn't sufficient.

Video Duration

For video placements, the validator also checks duration. A TikTok In-Feed video must be 5–60 seconds. YouTube bumper ads must be ≤6 seconds.

Platform Categories

Channels are grouped into categories for easier navigation:

Social — Facebook, Instagram, YouTube, TikTok, LinkedIn, X/Twitter, Pinterest, Snapchat, Reddit, Threads, Bluesky, WhatsApp, Telegram, Discord, Twitch
Display (GDN) — All standard IAB display sizes
Programmatic — DV360, The Trade Desk
CTV / OTT — Connected TV formats
Business — Google Business Profile, Quora

AI Resize

Resize uses Cloudinary URL-based transformations. The tool constructs the correct transform chain based on how different the target size is from your source image.

Strategy Selection

Scenario	Transform	What Happens
Close aspect ratio (within ±0.2)	c_fill,g_auto	Smart crop with AI gravity detection. Keeps the subject centered, trims edges minimally.
Different aspect ratio (beyond ±0.2)	c_pad,b_gen_fill	AI generative fill extends the background. No content is cropped — new background is generated to fill the required canvas.

Why generative fill matters Traditional resize either crops (you lose content) or stretches (the image distorts). Generative fill understands the image context and creates new, seamless background that matches the original. A product shot on a white table extends naturally to a wider canvas.

Resize Packages

Pre-built packages let you generate all sizes for a platform in one click:

GDN Essential

8 sizes

300×250, 336×280, 728×90, 300×600, 320×50, 160×600, 250×250, 200×200. Covers ~95% of available GDN impression volume.

Meta Complete

6 placements

Feed (1080×1080), Story/Reel (1080×1920), Landscape (1200×628), Carousel (1080×1080), Right Column (1200×628), Marketplace (1200×628). Every Meta placement covered.

TTD Complete

11 sizes

All standard Trade Desk display sizes including native, banner, and interstitial formats.

DV360 Complete

12 sizes

Full DV360 display size coverage including custom high-impact formats.

Display All (Universal)

15 sizes

All IAB standard display sizes. One click generates every standard banner, skyscraper, rectangle, and leaderboard.

Animate

Convert any still image into a short video ad using Google Veo 3.1 (image-to-video).

How It Works

Upload or select an already-validated image.
Choose a duration: 4s, 6s, 8s, or 10s.
Select a style: zoom, pan, parallax, cinematic, or let AI decide.
Click "Animate" — the image is sent to Google's Veo 3.1 model, which generates a video from the still frame.

Requirements

A Google AI Studio API key (get one free at aistudio.google.com/apikey)
Source image should be high quality — Veo performs best with sharp, well-composed images
Generation takes 30–90 seconds depending on duration and complexity

Veo 3.1 availability Veo 3.1 is available through the Google AI Studio free tier. No billing required. Rate limits apply (varies by region).

Platform Specs

Social Platforms

Platform	Placement	Size (px)	Ratio	Max File
Facebook	Feed	1080×1080	1:1	30 MB
Facebook	Story	1080×1920	9:16	30 MB
Facebook	Right Column	1200×628	1.91:1	30 MB
Instagram	Feed Square	1080×1080	1:1	30 MB
Instagram	Feed Portrait	1080×1350	4:5	30 MB
Instagram	Story / Reel	1080×1920	9:16	30 MB
YouTube	Display Companion	300×250	1.2:1	150 KB
YouTube	Overlay	480×70	6.86:1	150 KB
TikTok	In-Feed	1080×1920	9:16	500 MB
TikTok	TopView	1080×1920	9:16	500 MB
LinkedIn	Single Image	1200×627	1.91:1	5 MB
LinkedIn	Story	1080×1920	9:16	5 MB
X / Twitter	In-Stream	1200×675	16:9	5 MB
X / Twitter	Card	800×418	1.91:1	5 MB
Pinterest	Standard Pin	1000×1500	2:3	20 MB
Pinterest	Max-Width Pin	1000×2100	1:2.1	20 MB
Snapchat	Full Screen	1080×1920	9:16	5 MB

Google Display Network (GDN)

Name	Size (px)	Ratio	Usage
Medium Rectangle	300×250	1.2:1	Most common — text pages, forums
Large Rectangle	336×280	1.2:1	Larger in-content placement
Leaderboard	728×90	8.09:1	Top of page, below navigation
Half Page	300×600	1:2	High-impact sidebar placement
Mobile Leaderboard	320×50	6.4:1	Mobile header/footer
Wide Skyscraper	160×600	1:3.75	Sidebar tower
Square	250×250	1:1	Versatile square format
Small Square	200×200	1:1	Compact square
Banner	468×60	7.8:1	Classic banner
Large Leaderboard	970×90	10.78:1	Premium desktop header
Billboard	970×250	3.88:1	High-impact desktop top
Large Mobile Banner	320×100	3.2:1	Mobile interscroller

The Trade Desk (TTD)

Name	Size (px)	Max File
Medium Rectangle	300×250	200 KB
Leaderboard	728×90	200 KB
Wide Skyscraper	160×600	200 KB
Half Page	300×600	200 KB
Billboard	970×250	200 KB
Mobile Leaderboard	320×50	200 KB
Large Rectangle	336×280	200 KB
Large Leaderboard	970×90	200 KB
Mobile Banner	300×50	200 KB
Native	1200×627	1 MB
Native Square	1200×1200	1 MB

DV360

DV360 supports all standard GDN sizes plus additional high-impact formats. The validator checks against DV360-specific file size limits (typically 200 KB for standard display, 1 MB for native and rich media).

CTV / Connected TV

Format	Resolution	Ratio	Duration
Standard CTV	1920×1080	16:9	15s, 30s, 60s
4K CTV	3840×2160	16:9	15s, 30s, 60s
Companion Banner	300×250	1.2:1	—

MCP Server

The Ad Creative MCP server exposes validate, resize, and specs tools to any MCP-compatible AI client. Install it to validate and resize assets through natural language in Claude, Cursor, or any MCP host.

Installation

      bash
npm install -g ad-creative-mcp
    

Cursor Configuration

Add to your .cursor/mcp.json:

      json
{
  "mcpServers": {
    "ad-creative": {
      "command": "npx",
      "args": ["-y", "ad-creative-mcp"],
      "env": {
        "CLOUDINARY_CLOUD_NAME": "your-cloud-name",
        "CLOUDINARY_API_KEY": "your-api-key"
      }
    }
  }
}
    

Claude Desktop Configuration

Add to your claude_desktop_config.json:

      json
{
  "mcpServers": {
    "ad-creative": {
      "command": "npx",
      "args": ["-y", "ad-creative-mcp"],
      "env": {
        "CLOUDINARY_CLOUD_NAME": "your-cloud-name",
        "CLOUDINARY_API_KEY": "your-api-key"
      }
    }
  }
}
    

Claude Code

Add the MCP server via the CLI:

      bash
claude mcp add ad-creative -- npx -y ad-creative-mcp
    

Available MCP Tools

Tool	Description
validate_creative	Check an image/video against all platform specs. Returns compatibility matrix.
resize_creative	Generate a resized derivative using AI fill. Returns Cloudinary URL.
batch_resize	Resize to multiple sizes at once (e.g., entire GDN package).
get_platform_specs	Return specs for a specific platform or all platforms.

Example prompt "Validate my hero image at /images/hero.jpg against all GDN sizes and resize any that don't match."

Gemini Setup

Use the Creative Validator through Gemini CLI, Google AI Studio, or the Gemini REST API.

Gemini CLI

Install the Gemini CLI and configure your tool settings in ~/.gemini/settings.json:

      json
{
  "mcpServers": {
    "ad-creative": {
      "command": "npx",
      "args": ["-y", "ad-creative-mcp"],
      "env": {
        "CLOUDINARY_CLOUD_NAME": "your-cloud-name",
        "CLOUDINARY_API_KEY": "your-api-key"
      }
    }
  }
}
    

AI Studio Function Declarations

When using Gemini through AI Studio or the REST API, declare the validator as a function:

      json
{
  "name": "validate_creative",
  "description": "Validate an advertising creative asset against platform specs",
  "parameters": {
    "type": "object",
    "properties": {
      "image_url": {
        "type": "string",
        "description": "URL of the image to validate"
      },
      "platforms": {
        "type": "array",
        "items": { "type": "string" },
        "description": "Platforms to check (e.g., gdn, meta, tiktok). Omit for all."
      }
    },
    "required": ["image_url"]
  }
}
    

REST API Usage

Call the validator from Gemini's function-calling flow. The model will invoke the declared function when the user asks about creative sizing, and you handle the tool call by forwarding it to the validator API.

      bash
curl -X POST "https://generativelanguage.googleapis.com/v1/models/gemini-2.0-flash:generateContent" \
  -H "Content-Type: application/json" \
  -H "x-goog-api-key: YOUR_API_KEY" \
  -d '{
    "contents": [{"parts": [{"text": "Validate my banner image for GDN"}]}],
    "tools": [{"function_declarations": [{ ... }]}]
  }'
    

Claude Setup

Use the Creative Validator with Claude Desktop or Claude Code through the MCP server.

Claude Desktop

After adding the MCP config (see MCP Server above), restart Claude Desktop. The validator tools appear automatically. Try these prompts:

      example prompts
"Validate this image for all social media platforms"
"Resize my banner to all GDN sizes using AI fill"
"What are the specs for TikTok In-Feed ads?"
"Generate all Meta placements from my 1080x1080 image"
"Show me which platforms this 1200x628 image is compatible with"
    

Claude Code

After adding the MCP server with claude mcp add, use it directly in your terminal:

      bash
claude "Validate the hero image in /assets/hero.jpg against all ad platforms"
claude "Batch resize /assets/product.png to all GDN Essential sizes"
claude "What dimensions does LinkedIn require for sponsored content?"
    

Claude works best when specific Instead of "resize my image," try "resize /path/to/image.jpg to all GDN sizes using generative fill." Including the file path and target platform yields the best results.

Cloudinary Cowork

Cowork is a no-code way for non-technical team members to interact with the Creative Validator through a shared Cloudinary workspace.

How It Works

Share the workspace — Invite team members to your Cloudinary account with upload-only permissions.
Upload to a folder — Team members upload creative assets to a designated folder (e.g., /ad-creatives/incoming).
Auto-validate — A Cloudinary notification webhook triggers validation on upload. Results are written as structured metadata on the asset.
Auto-resize — If configured, derivatives are automatically generated using the correct transform chain and saved to platform-specific subfolders.

Requires Cloudinary Programmable Media plan Notification webhooks and auto-tagging require at least the Cloudinary Plus plan. The free tier supports manual validation and resize through the web app.

Folder Structure

      folder layout
/ad-creatives/
  incoming/          ← upload originals here
  validated/         ← auto-moved after validation
  resized/
    gdn/             ← 300x250, 728x90, etc.
    meta/            ← feed, story, carousel
    ttd/             ← TTD display sizes
    social/          ← LinkedIn, X, Pinterest
    

API Reference

POST /api/creative-tools

All operations go through a single endpoint. The action field determines what happens.

action: "validate"

      request
{
  "action": "validate",
  "image_url": "https://example.com/banner.jpg",
  "width": 1200,
  "height": 628
}
    

      response
{
  "source": { "width": 1200, "height": 628, "ratio": 1.91 },
  "results": {
    "facebook_feed": { "status": "resize", "target": "1080x1080", "strategy": "gen_fill" },
    "facebook_story": { "status": "resize", "target": "1080x1920", "strategy": "gen_fill" },
    "linkedin_single": { "status": "compatible", "target": "1200x627" },
    "gdn_leaderboard": { "status": "resize", "target": "728x90", "strategy": "fill" },
    ...
  },
  "summary": { "compatible": 12, "resize": 38, "incompatible": 3 }
}
    

action: "transform"

      request
{
  "action": "transform",
  "public_id": "ad-creatives/hero-banner",
  "cloud_name": "your-cloud",
  "target_width": 300,
  "target_height": 250,
  "strategy": "auto"
}
    

      response
{
  "url": "https://res.cloudinary.com/your-cloud/image/upload/c_pad,w_300,h_250,b_gen_fill/ad-creatives/hero-banner.jpg",
  "transform": "c_pad,w_300,h_250,b_gen_fill",
  "width": 300,
  "height": 250
}
    

action: "batch"

      request
{
  "action": "batch",
  "public_id": "ad-creatives/hero-banner",
  "cloud_name": "your-cloud",
  "package": "gdn_essential"
}
    

      response
{
  "package": "gdn_essential",
  "count": 8,
  "derivatives": [
    { "name": "Medium Rectangle", "size": "300x250", "url": "https://res.cloudinary.com/..." },
    { "name": "Large Rectangle", "size": "336x280", "url": "https://res.cloudinary.com/..." },
    { "name": "Leaderboard", "size": "728x90", "url": "https://res.cloudinary.com/..." },
    ...
  ]
}
    

action: "specs"

      request
{
  "action": "specs",
  "platform": "meta"
}
    

      response
{
  "platform": "meta",
  "placements": [
    { "name": "Feed", "width": 1080, "height": 1080, "ratio": "1:1", "maxFileSize": "30MB" },
    { "name": "Story", "width": 1080, "height": 1920, "ratio": "9:16", "maxFileSize": "30MB" },
    { "name": "Right Column", "width": 1200, "height": 628, "ratio": "1.91:1", "maxFileSize": "30MB" },
    ...
  ]
}
    

Cloudinary Setup

Cloudinary provides the image transformation engine. You need a free account to use resize and animate features.

Create a Free Account

Find Your Cloud Name

After signup, your dashboard shows your Cloud Name (e.g., dxyz123abc). This is the account identifier used in all transformation URLs.

Get Your API Key

Go to Settings → Access Keys. Copy the API Key (numeric string). You do not need the API Secret for client-side unsigned uploads.

Enable Unsigned Uploads

Go to Settings → Upload. Scroll to Upload presets and create an unsigned preset or enable the default one. This allows the browser app to upload images directly.

Enter in the App

Open the Creative Validator, click the Cloudinary settings panel, and paste your Cloud Name and API Key. They're stored in sessionStorage — never sent anywhere except Cloudinary.

Free tier limits Cloudinary's free plan includes 25 credits/month. A single resize costs ~1 credit. A GDN Essential batch (8 sizes) costs ~8 credits. Generative fill transforms cost ~2 credits each. For heavy usage, the Plus plan starts at $89/month.

FAQ

Do I need an account to validate?

No. Validation runs entirely in your browser. Upload an image, see results instantly. No account, no API key, no signup.

Do I need a Cloudinary account to resize?

Yes. Resize uses Cloudinary's transformation API. A free account works for moderate usage (25 credits/month).

What image formats are supported?

JPG, PNG, WebP, GIF for images. MP4, MOV for video. The validator reads dimensions client-side for all formats.

Is my image uploaded anywhere during validation?

No. During validation, your image never leaves your browser. Dimensions are read locally using the browser's Image/Video API. Only during resize is the image sent to Cloudinary.

How accurate is generative fill?

Cloudinary's AI generative fill (powered by their ML pipeline) produces seamless extensions for most photographic content. It works best with images that have consistent backgrounds (product shots, landscapes, office scenes). Complex patterns or text near edges may produce artifacts.

Can I use this with the Cloudinary MCP server?

Yes. The Ad Creative MCP server works alongside the official Cloudinary MCP server. You can use either or both. The Ad Creative MCP adds platform-specific validation logic that the generic Cloudinary MCP doesn't have.

What's the difference between c_fill and c_pad?

c_fill,g_auto crops to fit (with smart gravity to keep the subject centered). c_pad,b_gen_fill adds AI-generated padding without cropping. The tool picks automatically based on how different the target ratio is from the source.

Can I animate a video?

Not currently. The animate feature is image-to-video only (using Veo 3.1). Video-to-video editing is on the roadmap.

Is there a batch upload option?

The web app handles one image at a time. For batch processing, use the MCP server or API endpoint with scripts.

Where are my Cloudinary credentials stored?

In your browser's sessionStorage. They are scoped to the tab, never persisted to disk, never sent to any server other than Cloudinary, and are destroyed when you close the tab.