Character swap vs. full ad localization: what's the difference in cost and effort?

A character swap changes one asset layer and costs about $30 (115 credits) and 30 minutes per ad once the workflow is set. Full localization regenerates character sheets, location sheets, voiceover, and on-screen UI text — roughly $70–$150 and 1–2.5 hours per market, about 5x the credits per ad.

The cost gap comes directly from scope: a swap regenerates only the character or product layer while every other locked asset — edit structure, shot list, music, pacing — carries over untouched, whereas full localization regenerates four asset tiers for the new market. Both workflows run inside the invideo agent, which holds the original ad's locked assets in project context so unchanged layers never get regenerated or re-paid for.

The numbers, side by side. These are documented production actuals, and cost benchmarks include rejected generations (one production rejected ~85% of all clips and still landed at these figures):

Localization costs vary by team and ad complexity: one production ran $145 per localized ad at 570 credits; another localized three winning ads into two markets (Japan and Spain) for $425 total — about $70 per ad and 2.5 hours each; a multi-agent batch run landed around $150 per new market variant. Treat $70–$150 as the honest range.

What a swap involves. Upload reference images of the new outfit or product — front, side, back, and a close-up — and the invideo agent analyzes the original video, splits it into segments (it detects outfit changes on its own), regenerates each segment with the new asset, and you lock the best option per segment and stitch. The first swap takes about 2 hours to establish the workflow; every one after that takes about 30 minutes, which is what makes 12 swaps in an 8-hour day possible.

What full localization involves. Ask the invideo agent to analyze the original ad first — it returns a change plan listing what stays, what changes, and an estimated generation breakdown, so you know approximate spend before committing credits. Then lock tiers in sequence: new character and location reference sheets (about 4 per market, ~5 minutes each with GPT-Image-2), roughly 13 regenerated video clips mapped against the original's preserved shot beats (one production held 7 beats structurally identical), around 5 translated and regenerated app UI screens, and a translated voiceover with lip-sync in the target language. The edit order, music bed, and pacing stay constant — only cast, location, voice, and on-screen text change. The first market takes about 2 hours; each additional market about 1 hour, and you can batch markets in one pass by naming character reference files after their target language (Spanish.jpg, French.png) — one production scaled to 9 markets this way.

Which one to run. Use a swap when the creative already works for the audience and only the product or character layer changes — same language, same cultural register. Use full localization when the market needs new casting, a new language, and adapted copy: one localized version rewrote the on-screen CTA into Japanese ('up to 45% off on first purchase') alongside the visual changes. That's closer to what the localization industry calls transcreation — adapting the creative intent, not translating the words — and it's why localization costs 5x the credits: you're rebuilding the cultural layer, not repainting one asset. Combined, one senior creative running both workflows can produce 8–10 variations of a single winning ad per day.

Watch some of these to see what works for you:

I rejected about 85% of all clips. Each product swap cost me about 115 credits, which is roughly around $30. And each localization ad cost me about 570 credits, which comes to about $145.

— invideo's creative team