Guides

Practical Ad Creative Guides

Direct answers to the questions ad production actually asks — sizes, safe zones, text handling, automation.

RegenerationMasterSALE 30%regenerateper size →SALE 30%SAIE 3O%SALE 30%SLAE 30%… N chances to breakPreservationMasterSALE 30%preserve text,rebuild background →SALE 30%SALE 30%SALE 30%SALE 30%… original intact
Regenerating per size multiplies the chances of garbled text by the number of sizes. Preservation never redraws the text, so there is no path to breakage

Fixing Garbled Korean Text in AI Ads — a 3-Step Procedure

A directly applicable 3-step fix for garbled Korean in AI-generated ads: generation rules, preservation-based size adaptation, and a review order that catches what matters first.

1440×1970250px340px

Story

1268×1306358px896px

Reels

Same 1440×2560, different safe areas: Stories (left) vs Reels (right) — Reels lose up to 35% at the bottom, leaving roughly a square

Instagram Story & Reels Ads: Where Text Is Actually Safe — in Pixels

Stories lose the top 14% and bottom 20% to UI; Reels lose up to 35% at the bottom. Safe-area pixel values at 1440×2560 and placement rules for text and CTAs.

One master→… text preserved
Not rebuilt per size — one master fans out to dozens in parallel, with text and logo kept as original pixels.

Bulk Banner Resizing — Cutting Size Adaptation to 10 Minutes

The right structure is "one master → batch adaptation", not "remake per size". The four requirements for bulk resize automation and an operations checklist.

Manual round-tripFrame→Export→External tool→Re-importPluginMaster frame→Plugin→N sizes (in Figma)
Drop the export→tool→re-import round-trip; a plugin expands one master frame into N sizes inside Figma.

Automating Banner Resize in Figma — Size Adaptation Without Leaving Your Design Tool

Where manual Figma adaptation breaks down (constraints only stretch), and the plugin flow: pick a master frame, choose placements, get platform-spec results.

"Remove the badge" (name only)Full regen — other elements change0/12"Remove the top-left red badge, fill with background, keep the rest"Only the badge removed — rest intact9/9
Measured: a name-only instruction applied 0/12; one with position, look and after-state applied 9/9 (100%).

Editing One Element with AI — Without Everything Else Changing

Why models repaint the whole canvas even for a small edit, and the instruction pattern that took apply-rate from 0% to 100% in our tests: location, appearance, and the after-state.

Master→1:11:1 Feed4:54:5 Vertical9:169:16 Story·Reels1.91:11.91:1 Link
The default Meta delivery set — one master into a 4-ratio set, covering every auto-placement.

Every Image Size You Need for Meta Ads — Feed, Story, Reels, Carousel

The image sizes a Meta campaign actually needs, by placement: recommended dimensions, ratios, safe zones and carousel rules on one page.

Creative Adaptation, Versioning, Variations — Ad Production Terms, Untangled

Resize, size adaptation, creative versioning, variations, creative refresh — what each term actually means in practice, and which ones automation handles best.

Canva Magic ResizeEditable Canva design only→A finished JPG breaks the premiseImageFactoryFinished image (any tool)→Auto-adapts to spec & safe zone
The key difference is the input. Canva assumes an editable design; ImageFactory takes a finished image (any tool) and adapts it to ad specs.

Canva Magic Resize vs ImageFactory — Which Fits Ad Size Adaptation?

The structural difference: Canva Magic Resize moves an editable Canva design to another canvas; ImageFactory adapts a finished image to platform specs with safe zones applied.

Korean Banner Automation Tools Compared — VCAT, VariAid, ImageFactory

Three different starting points for the same problem: URL-based generation, spec-template tools, and finished-image adaptation. How to choose by situation.

Try it on your own creative

One master creative becomes every placement size — with the text preserved exactly. 14-day free trial.

Get started free