Research wiki · synthetic NCII
Queries like ai undress, undress app, nudify and face swap porn pull in millions of searches every month. This page documents what those services actually are: the market data, the newsroom investigations into who runs them, the abuse cases that reached police and courts, and the protocol that works when someone is targeted.
An AI undress app takes an ordinary clothed photo of a real person and generates a fabricated nude or sexualized image of them. The model does not recover anything hidden: it hallucinates a body that was never photographed. What stays real is the identity, the face, and that is the abuse. Searches for ai undressing, ai undresser, undresser ai, undress ai app or ai undress app all describe this same function.
Two adjacent categories share the outcome. Nudify services (queries: nudify, nudify ai, nudify app) are the same undress function under newer branding, a line that runs from the 2019 DeepNude desktop app to today's web tools. Face swap porn (also searched as porn face swap, faceswap porn, deepfake porn) grafts a real face onto existing explicit material instead of inventing a body. Researchers and prosecutors group all of it as synthetic non consensual intimate imagery, shortened to synthetic NCII.
The category operates at industrial scale. Independent measurements converge on the same picture: tens of millions of monthly visits, and targets that are overwhelmingly women.
| Dataset | Year | Finding |
|---|---|---|
| Graphika, synthetic NCII report | 2023 | 34 undress providers, about 24 million visits in September 2023. Spam links promoting them on X and Reddit grew more than 2,000 percent across the year. |
| Security Hero, State of Deepfakes | 2023 | Roughly 98 percent of deepfake videos online are pornographic; about 99 percent of people targeted are women. |
| Sensity, Telegram bot census | 2020 | A DeepNude style bot ecosystem on Telegram had produced over 100,000 fake nude images, mostly of private individuals. |
| San Francisco City Attorney, complaint | 2024 | 16 undress sites received more than 200 million visits in the first half of 2024; operators described as anonymous and unaccountable. |
| Internet Watch Foundation | 2024 | Rising reports of AI generated child sexual abuse imagery, including material derived from ordinary photos of real children. |
Four cases that reached police, prosecutors or federal court. Each entry links the technology to documented human harm.
More than 20 girls found fabricated nude images of themselves circulating in local WhatsApp groups. El Pais reported the ClothOff watermark on one image. The case went to juvenile prosecutors; some victims were also blackmailed. A mother told the Guardian the image of her daughter was realistic enough to pass for real.
Boys generated fake nudes of female classmates with an AI tool and shared them in a group chat. The Guardian later identified the same app behind both this case and Almendralejo. The school had no policy for a crime that barely existed a year before; students walked out in protest.
Around 50 girls were targeted with fabricated explicit images built from their social media photos. Victoria Police arrested a teenage boy. Australia's eSafety Commissioner now treats synthetic image based abuse as a priority enforcement area.
AI generated explicit images of Taylor Swift spread across X in late January 2024; one post passed 45 million views before removal, and the platform briefly blocked searches of her name. Within days, the White House and members of Congress were commenting on synthetic NCII.
A federal complaint filed on October 16, 2025 targets AI/Robotics Venture Strategy 3 Ltd., a British Virgin Islands entity operating the ClothOff network, and alleges the production and sale of non consensual sexual images of children and adults. It is the first case in which a victim goes after the operating company and its alleged owners directly.
Four newsrooms and two courts have mapped who runs the leading undress services. The pattern is consistent: deliberate anonymity, masked money, no accountability.
These are the queries that route people into this category, grouped by what the searcher is actually looking for. Each group shares the same documented mechanics and risks.
Using these services is not a grey zone hobby. The documented exposure runs in four directions.
Charges masked as game purchases, no refunds, anonymous beneficiaries. The business model counts on users being too embarrassed to dispute.
Uploads land on servers run by operators with fake addresses and fake executives. One of their databases has already been found exposed on the open internet.
Juvenile prosecutions in Spain, an arrest in Australia, school discipline in New Jersey, and victims now suing operators in federal court. Creating or sharing the output is a crime in a growing list of jurisdictions.
If the subject is under 18, the output is child sexual abuse material in the US, UK, EU and Australia, synthetic or not. It is the most prosecuted category of online crime.
If someone ran your photos through an undress AI service, act in this order. The tools named are free.
Screenshots, URLs, usernames, dates. Do it before reporting, because removal deletes the evidence along with the content.
StopNCII.org (18+) or Take It Down by NCMEC (under 18 in the original photo) hash the image on your device and share only the hash with partner platforms, including Meta and TikTok.
Every major platform has a dedicated category for non consensual intimate imagery, synthetic included.
File Google's non consensual explicit imagery removal request so copies stop surfacing in Search.
Police, with the preserved evidence. US support: Cyber Civil Rights Initiative helpline, 1-844-878-2274. The 2025 New Jersey case shows civil action against operators is now real.
An AI undress app is a service that fabricates a fake nude of a real person from an ordinary clothed photo. The body in the output is invented by a model, but the face is real and used without consent, which is why prosecutors treat the output as image based abuse, or as child sexual abuse material when the subject is under 18.
In a growing number of countries and US states, creating or sharing fabricated sexual images of real people is a crime. Documented consequences include juvenile prosecutions in Spain, an arrest in Australia, student discipline in New Jersey, and civil lawsuits against operators in San Francisco and New Jersey.
An undress app invents a synthetic body under a real face from a clothed photo. Face swap porn grafts a real face onto existing explicit photos or videos. Both attach a real identity to sexual content the person never made, and both sit inside what researchers call synthetic non consensual intimate imagery.
No. Generators synthesize pixels; they do not reveal anything real about a person. The abuse starts when a real, identifiable face is attached to the output without consent. At that point the legal and human harm is the same as with any other fabricated intimate image.
Documented reporting on this category says no. Bellingcat found a leading service masking charges as video game purchases, CBS 60 Minutes found a listed office address that belonged to an unrelated company, and Der Spiegel found an operator database exposed on the open internet. There is no verifiable privacy or refund policy behind anonymous operators.
DeepNude, the $50 desktop app that started the undress category in 2019, was shut down by its own creator less than a day after Vice Motherboard exposed it, and GitHub later removed its repositories. MrDeepFakes, the largest face swap porn hub, closed in May 2025 after losing critical service providers.
Save evidence first: screenshots, links, usernames, dates. Then hash the image with StopNCII.org (adults) or Take It Down from NCMEC (minors in the original photo) to block reuploads, report to the hosting platform, file a Google Search removal request, and go to police with the saved evidence.
They can be pointed at any photo, and documented school cases in Spain, the US, Australia and South Korea show they have been. The output is treated as child sexual abuse material in the US, UK, EU and Australia even though it is synthetic, which makes it one of the most prosecuted categories of online crime.
They are cheap to run and profitable. Graphika counted about 24 million visits to 34 services in one month of 2023, and a San Francisco filing cites more than 200 million visits to 16 sites in the first half of 2024. When hosts or payment processors cut one domain off, the operators relaunch under another.
This page is maintained by the aiundress.wiki Research Desk, an independent documentation project on synthetic non consensual intimate imagery. Every claim above cites a dated public source: a court filing, a police statement, a peer reviewed or investigative dataset, or named newsroom reporting. The page is informational, is not legal advice, and names no service for any purpose other than documenting published findings.
Editorial review: Digital Violence Studies Group, Universidad Nacional de San Luis (UNSL), Argentina. Founded in 1973, with an institutional legacy reaching back to 1939, UNSL is a public national university whose research groups work on technology and society, including digital violence against women and minors.