Retouching by sentence instead of by selection
Most photo editors ask you to select before they will erase. You drag a lasso, feather the edge, paint a mask, then discover the removal left a smudge where a shoulder used to be. People Remover swaps that sequence for a description: upload the picture, say who should disappear, download the result. The instruction replaces the mask.
That sounds like a small convenience until you are working through a batch. A museum shot with a stranger in the corner, a team photo where one person has left the company, a storefront image with a passer-by crossing the frame. Each one used to be a minute of careful selection. Now it is a sentence, and the sentence is reusable.
What the tool actually does
According to the product page, the workflow is upload, describe, download. There is no brush and no mask layer. Descriptions can be positional or descriptive, so "the man on the left" and "the woman in the red coat" are both valid ways to point at a subject. Because one instruction counts as one edit regardless of how many people it names, clearing a small crowd is a single pass rather than one operation per person.
The same page lists JPG, PNG, WebP and iPhone HEIC as accepted inputs, with a file size ceiling on uploads, and JPG, PNG or WebP as download formats without a watermark. Output is offered at several sizes, with the larger ones sitting behind paid plans; the free path gives a single smaller export after signing in. Treat those numbers as the vendor's own statement rather than a measurement.
A short working checklist
Before uploading anything, write the instruction as if you were briefing a colleague: who, where, and one distinguishing detail. "Remove the cyclist" is weaker than "remove the cyclist in the yellow jacket on the right". Then check three things after the download:
1. The removed area itself, at full size rather than in a thumbnail.
2. The surround, because a full redraw can shift small details such as text on a sign or the edge of a table.
3. The subject you kept, especially hands and hair where generated fills tend to show.
If the wrong person went, do not add more words to the same prompt. Rewrite it from the position rather than the appearance, or crop tighter so there is only one candidate for the instruction to match. Keeping a note of which phrasings worked for a given scene type saves the next round.
Where it fits in a real queue
The obvious home for a tool like People Remover is a retouch queue that is not worth a designer's afternoon: listings photography, event recaps, travel images destined for a blog header. It is a poor fit for anything that requires pixel-exact reconstruction of architecture or legible signage, and a worse fit for editorial work where the presence of a person is the point.
There is also a consent question that no tool answers for you. Removing a stranger from your own holiday photo is one thing; removing a credited colleague from a press image is another. Decide the policy before the folder arrives.
Bottom line
If your recurring problem is "there is a person in this photo and I do not want to spend ten minutes selecting them", the describe-and-download model is the right shape. Start with one image you can afford to redo, judge the surround rather than only the hole, and keep the phrasing that worked. You can see the current input limits and export sizes on the People Remover product page before committing a batch to it.
