Body editing is usually discussed as one category, and it is really two. One direction adds weight; the other removes it. They have never been treated the same way — not by the platforms that shipped them, not by the comment sections that received them, and not by the tools that offer them now. Something like vofy puts a fat filter in front of a user as a preview they can keep or discard, which places it in the first direction. Which direction a product runs in, and what that quietly says about the person using it, is what this piece is about.

A see-saw is the wrong picture for these two effects, because a see-saw implies balance. The record of the last year does not show balance. It shows a pair of products doing opposite things to the same photograph, and only one of them being asked to justify itself.

One app shipped both directions

CapCut, the editing app owned by ByteDance and distributed alongside TikTok, carried an effect that added weight to a face. It became known as the chubby filter, and it went viral in March 2025; Cybernews and the Daily Dot both traced its spread back to CapCut, and reporting in the New York Post and The Independent placed the peak of the reaction in the same month.

On 22 March 2025 the effect was removed. TikTok confirmed the change to the BBC and to the New York Post, and the removal applied to CapCut, the app the filter had actually lived in.

On the evidence collected at the time, that was a defensible decision. The Independent reported users describing the effect as a way to mock larger people and to install thinness as the only acceptable target. The Daily Dot documented derogatory comments landing on the altered clips, where the joke was the body rather than the transformation. The New Daily reported accounts applying the effect to photographs of people who had not agreed to it.

The other direction was not removed. The New Daily's write-up noted that the "skinny filter" — the reverse version, built to take weight off rather than add it — drew far less controversy. No campaign formed around it, so no spokesperson was asked to explain it.

Two effects, one app, opposite signs on the same transformation. Only one of them ended in a moderation decision.

The two directions are not mirror images

The natural assumption is that these are the same product with the sign flipped, and that if one of them is a problem, the other is the same problem facing the other way. The reporting does not support that.

Start with the captions. The weight-adding videos that drew the backlash were framed as exposure — a slim clip first, the altered version second, with the result treated as the embarrassing part. Underneath them, the reactions treated the body as the punchline, which is why a teenage health app called Luna publicly criticised the effect for reinforcing body shame, as covered by Talking Pretty, and why a creator with 66,000 followers told interviewers the filter made her want to scream (via pyLessons and Yahoo). Betches documented a creator, Becki Jones, leaving TikTok after sustained weight-based bullying.

Then read what people wrote under the opposite direction. Grazia Daily quoted users posting the slimming version with a caption that went roughly: look at what I would look like if I lost weight — motivation. Not a punchline. A destination.

Betches drew the conclusion that follows from that contrast: being told you are too thin tends to attract concern, while being told you are too heavy tends to attract mockery. Grazia's own verdict on the trend was blunter — "mockery disguised as a joke." Applied to the videos, that description is fair. Applied to the capability itself, it is not, and that gap between the videos and the tool is where most of this debate has quietly fallen apart.



Only one of them reads as a goal

Stated plainly: an effect that removes weight reads as an improvement, and an effect that adds weight reads as a warning. Both are opinions about which body is the destination.

That is not an argument that the weight-adding direction is harmless. It is not, and the documented harms are specific rather than vague — non-consensual edits, mocking replies, a body-image app intervening publicly, creators describing real distress. A platform looking at that list and withdrawing the effect made a call that holds up.

But the call was not neutrality. Removing one direction while leaving the other is itself a position: it treats the thinner version as the fix and the heavier version as the insult. Anyone who has read both comment sections already knows which of the two the internet treats as a compliment, and which one it treats as a verdict.

That is the asymmetry the 2025 coverage never quite said out loud. A tool offering both directions is handing over a choice. A tool offering only the slimming direction is not handing over a choice — it is repeating the comment section, in software, without ever having to write a sentence about it.

Why the choice now belongs to whoever builds the tool

In 2025 the direction was decided by one company's moderation team. In 2026 it is decided by product copy, because the effect stopped being a store listing and became something the underlying models simply do.

The category normalised fast. Consumer roundups published in January, March and June 2026 (CheckMyFit, GainFrame) treat AI body editors as an ordinary thing to list, which is a reasonable measure of how unremarkable the capability has become. A general-purpose tool does not need anyone's approval to run the heavier direction, and it cannot be withdrawn from a shelf it was never on.

At the same time the surrounding environment stopped being forgiving. A swimwear campaign that digitally altered a model's body drew the #LandsEndFail backlash and roughly five million views inside twenty-four hours, as reported by The Tech Edvocate. The Atlantic's August 2026 piece on "algorithmic dysmorphia" argued the internet has become more image-centred than at any earlier point. Forbes reported in February 2026 on a split inside the body-positivity movement itself. None of those stories is about filters. All of them are about how quickly an edited body now reads as a claim about the person in it.

So the design question has moved. It is no longer whether the heavier direction should be permitted, because no store gets a vote on a capability that lives in a model. The question is what the product says while it hands the effect over — and the answer is now visible in the interface rather than buried in a policy document.



The framing on vofy's own page is a useful example because it is unusually explicit: the heavier result is described as a preview, the edits are meant to stay self-directed, prank versions are asked to stay clearly fictional, and users are told to check the output for background warping and identity stability before sharing it. That is a description of intended use written into the interface, and it exists because the page's direction is the one that needs the sentence.

The mirror image of that is the part worth noticing. A tool that offers only the slimming direction never has to write anything of the kind, and not needing the sentence is not the same as not needing the thought.

The takeaway

Two directions of body editing exist, and they have never been weighed equally. The weight-adding one was withdrawn after a backlash that was documented, specific and justified. The weight-removing one was left alone, because it reads as encouragement rather than insult — and anything that offers only that version is repeating the encouragement whether it intends to or not.

The honest reading is not that the heavier effect is innocent. It is that the direction is a decision, that the decision now sits with whoever builds the product rather than with a review team, and that describing the two effects as mirror images hides the decision instead of making it.