A文ListingTranslate
BLOG · July 12, 2026

How to Update Text on Product Photos Your Supplier Sent as JPGs

Your supplier sends the product photos, and they look great: the hero on white, the infographics with the specs called out, the badge that says the bottle is 12 oz. Then something on one of them is wrong. The size shipped as 10 oz, not 12. The “BPA-free” claim needs to change. A callout names a certification the product no longer carries. You open the file to fix the one line, and you hit the wall every seller eventually hits: what you have is a finished JPG, and the editable version that made it lives on the supplier’s machine, in a factory’s marketing folder, or nowhere you can reach.

This is the specific bind sellers are in that a design agency isn’t. You didn’t build these images, so you were never handed the layered source. The words are baked into the picture, the fonts are whatever the supplier’s designer happened to use, and the only copy you control is the one already exported. This post is about fixing the wrong words anyway, straight from the flattened file, without chasing anyone for a PSD that may not exist.

Why a supplier’s JPG can’t be edited on your side

A product photo isn’t a document, and a supplier’s export is the most locked-down version of that problem. Once the image is saved as a JPG or PNG, the text stops being text: it becomes pixels, fused into the same flat grid as the product and the background. There’s no cursor to click into, no layer to select, no font field to change. The “12 oz” on the badge isn’t a value you can edit; it’s an arrangement of coloured dots that happens to read as a number.

With work you commissioned yourself, you might at least ask for the source file back. With supplier imagery you usually can’t, for reasons that stack up:

So the file you can actually touch is the finished photo, and traditionally that meant the fix was somehow harder than making the image in the first place.

The two bad options sellers usually pick

Faced with a flattened supplier JPG, most sellers reach for one of two fixes, and both have a tell.

The first is to paste a text box over the wrong words. You white out “12 oz”, type “10 oz” on top, and hope. It never quite lands: the substitute font is a shade too heavy, the letter spacing is off, the colour is a hair different from the original, and the patch sits slightly proud of the artwork around it. On a hero infographic that a buyer zooms into, that mismatched box is exactly the kind of detail that makes a listing feel amateur, which is the opposite of what a supplier’s polished photo was doing for you.

The second is to brief a designer to rebuild the graphic. That gets you a clean result, but it’s slow and out of all proportion to the change. You’re paying someone to reconstruct a layout they didn’t design, source the fonts, re-type a single line and re-export, an hour or more of skilled work and a real invoice, all so one word can read differently. For a catalogue of any size, doing that every time a spec or claim shifts is how corrections quietly never happen, and wrong images stay live because fixing them costs more than living with them.

The fix: regenerate the photo with the words corrected

There’s a third path that skips the rebuild entirely. Instead of treating the supplier’s JPG as a project to reconstruct, you treat it as a photo to edit. You hand over the finished file, name the exact text that’s wrong and exactly what it should say, and the image is remade with only those words changed: same typeface, same colour, same position, everything else left as the supplier shot it.

That works precisely because you don’t need the source file, the fonts or the original assets. The correction happens on the flattened image itself. Quoting “12 oz” and “10 oz”, or the old claim and the new one, is enough to get the badge back reading correctly in its own type, with the product, the lighting and the surrounding callouts untouched. It’s the same in-place swap whether you’re correcting a spec, updating a claim you can no longer make, or changing a size that shipped differently from the artwork. The change-text-in-a-product-photo tool is built around exactly this case: the supplier’s finished file, a wrong line on it, and no layered original anywhere on your side. If the wrong line is a price or a promo rather than a spec, the price and promo swap handles that same operation on a number or an offer.

A quick honesty note, because this is remaking a photo rather than editing a document: always look at the result at full size, on the actual words you cared about, before you use it. Very small or very dense text, and heavily stylised script, are the cases most likely to need a second try. That’s the whole reason to check the finished image the way you’d check any file before it goes on a listing.

Doing it across a catalogue, and across marketplaces

One wrong badge is one image. The reason this approach matters to sellers is that the problem is rarely just one image. A supplier updates a formulation and half your infographics now name the wrong ingredient. A compliance change ripples across every SKU that carried the old claim. A pack size changes and the callout is wrong on the hero of a dozen listings at once. Handling that as a stack of finished files, each corrected in place, is what keeps it from becoming a design project you can’t afford to run.

The clearest version of this is expanding to a new country. When you take a listing from amazon.com to amazon.de or amazon.co.jp, the wrong words aren’t misspellings, they’re English words on an infographic that a German or Japanese buyer can’t read, and you still don’t have the layered file. The same in-place mechanism translates the text inside the image instead of correcting it: the size chart, the “what’s in the box” callout and the benefit annotations move into the target language with the layout intact. The Amazon workflow is the natural home for that: you point it at a listing, its existing gallery loads, you pick the images that carry selling text and the marketplaces you’re entering, and each one comes back with the text localised in place, every marketplace variant produced independently from your original image so a detail-heavy German version never bleeds into the French one.

One thing to be clear about, because it matters most for sellers: none of this touches your live listing, your store or your Seller Central account. It never connects to a marketplace at all. What you get back is a set of finished image files, named per marketplace in the case of the Amazon flow, and you upload them yourself, attaching each set to the right listing exactly as you would with anything a designer made for you. The tool hands you corrected pictures; where they go is entirely your call.

Start with your worst image

The way to try this is not to plan a whole catalogue sweep. Take the single supplier photo with the wrong word on it, the one you’ve been meaning to fix since it went live, and correct that. If it’s a spec, a claim or a callout, the product-photo tool is the place to start; the deeper background on working from a flattened file with no original is in the guide to editing text in an image without the source file.

The cost model is built for exactly that kind of low-stakes check. Your first image is free, sign in, no card required, so you can fix one real badge and judge the result before committing to anything. After that it’s one credit per finished image, and a failed image costs nothing, so a dense label that needs a second try never costs you extra. You only ever pay for a corrected photo you’ve looked at and decided to keep. Fix the one wrong word first, then check pricing when you’re ready to work through the rest of the catalogue.