Remove a Deepfake of You: Help Within the Hour (US)

Where it is right now

On Instagram, TikTok, X, Snapchat or Facebook. We report it under the platform's own rule on intimate images shared without consent, or on impersonation when the fake is not sexual. Where the federal 48-hour rule applies, we cite it, word for word. Then we check every day for reposts and report each one.

On a porn site. Most of these sites have a removal form they do not advertise. We find it, file your request with dated proof, and if the site stays silent, we go to the company that hosts it.

In Google results. Google removes fake sexual images of a named person from its search results on request, and a representative can file for you. We file, then follow up until the result is gone. Same with Bing.

In a private group, a Discord server or a Telegram channel. We report to the service's trust and safety team with the exact message links, and we ask the people you trust to send us what they see, so no copy gets missed.

How it works

Hour 1. You fill in a short form with the links and you pay. Your dedicated handler reads it and writes back within the hour, 8 am to 10 pm Eastern. If a link is missing, we tell you exactly what to send. Nothing else is on you.

Hour 2. The first reports go out, starting with what hurts most: the post your contacts can see, then the most viewed copies, then Google.

Days 1 to 7. Every day we search for new copies, report them, chase every report that got no answer, and push back on every refusal. Every day you get a short update by email. Our monitoring software, Pavois 4, looks for copies of the fake while we deal with the platforms.

Day 7. You get a report and an evidence file: each link, the date we captured it, who we reported it to, what they answered. Dated proof you can hand to the police or to any attorney you choose.

Start my emergency response

What the law says

In short: since May 19, 2026, a covered platform has 48 hours to take a sexual deepfake of you down once it gets a valid request. Our job is to make sure the clock starts, and to be there at hour 49.

The TAKE IT DOWN Act (Public Law 119-12, signed May 19, 2025) makes it a federal crime to knowingly publish an intimate image of a real person without consent, including an image made with AI. Since May 19, 2026, covered platforms must remove such an image within 48 hours of a valid request from the person or someone authorized to act for them, and make reasonable efforts to remove identical copies. The Federal Trade Commission enforces it, with civil penalties above $53,000 per violation, and takes reports at TakeItDown.ftc.gov. We write your requests in that form and cite the deadline.

Your state adds its own law. Criminal: California Civil Code 1708.86, Texas Penal Code 21.165 (extended to still images on September 1, 2025), New York Penal Law 245.15, Virginia Code 18.2-386.2, Florida Statutes 836.13, Pennsylvania 18 Pa.C.S. 3131, Rhode Island 11-64-3, South Carolina 16-15-332. Civil, so you can sue the person who made or shared it: Michigan's Protection from Intimate Deep Fakes Act (since August 26, 2025), California Civil Code 1708.86 (up to $250,000 since January 1, 2026), New York Civil Rights Law 52-c. All the laws, in plain English. General information, not legal advice.

What we will not do

We do not take cases involving anyone under 18. Use Take It Down by NCMEC, which is free, and report to the CyberTipline. We do not handle blackmail or sextortion. Do not pay, stop replying, keep the messages, and report to the FBI at ic3.gov. If you are in danger, call 911. If you are in crisis, call or text 988.

We have no back door into any platform. We do not file lawsuits, give legal advice, contact the person who made the fake, or recover money. We never ask for the image itself, a scan of your face, an ID document or a password. A link is enough.

Price

$299. One time. Charged in US dollars. No subscription, nothing renews, no upsell call.

Included: a dedicated handler within the hour, first reports within two hours, seven days of daily reports, searches and follow-ups, the day-7 report and the evidence file.

If our first reply is late, we refund the full amount on request and keep working anyway. If there is nothing we can report, we tell you within the hour and refund the full amount. You can cancel within 14 days and pay only for the days already worked.

What we can show

In October 2026, the first fake account we handled under Emergency Response, an Instagram account impersonating a performer, was taken down during the seven days.

Our monitoring software, Pavois 4, checked 146 fashion and beauty brands in October 2026 and found 78 fake stores using their names, 47 brands hit, 87 percent of the fakes created since May 2026. The figures are open data: DOI 10.5281/zenodo.23654418, also on data.gouv.fr.

Our founder writes a column for Journal du Net, a French business daily.

Who we are

SOS Deepfake is a French company based in Marseille. We work in English, by email, with clients in the United States and anywhere else. Everything is in writing, so everything is proof. We built this service for people who find a fake of themselves and get nothing but silence from the platform. Other tools ask for a scan of your face to search for it. We ask for the links. We know the US reporting routes, the federal 48-hour rule and the Google removal process, and we use them every day. Emergency Response runs from 8 am to 10 pm Eastern Time, seven days a week. You can send the form at any hour. A request that comes in overnight is read first thing at 8 am.

AI makes the fake. We protect what's real.

Questions people ask us

How do I remove a deepfake of me from the internet?

Four moves, in this order. Save the link to every page where it appears. Report it on each platform as an intimate image shared without consent, or as impersonation if it is not sexual. Ask Google to pull it from search results. Then follow up every day until it is gone. If you would rather open a case and get a dedicated handler within the hour, that is exactly what Emergency Response is.

How much does deepfake removal cost in the US?

$299, paid once, in US dollars. That buys seven days of work: a dedicated handler within the hour, first reports within two hours, daily searches and follow-ups, and a dated evidence file at the end. No subscription, nothing renews. Most competing services bill you monthly, by the year, or quote a price after a sales call.

How long does it take to get a deepfake taken down?

We reply within the hour, 8 am to 10 pm Eastern, and the first reports go out within two hours. Under the TAKE IT DOWN Act, a covered platform has 48 hours after a valid request to remove an intimate image. Google and porn sites take days, not hours. That is why we chase every report, every day, for seven days.

Is making a deepfake of me illegal in the United States?

If it is sexual, publishing it without your consent is a federal crime under the TAKE IT DOWN Act, and most states have their own law on top. Michigan, California and New York also let you sue the person who made or shared it. If it is not sexual, state laws on impersonation and the right to your own likeness may apply. The DEFIANCE Act, which would add a federal right to sue, passed the Senate on January 13, 2026 and is still waiting in the House as of October 27, 2026. General information, not legal advice.

Can you remove a deepfake from Google?

Yes. Google removes fake sexual images of an identifiable person from its search results on request, and a representative can file for you. Pulling it from Google hides the result, it does not delete the page. So we report the page itself at the same time.

What if the platform refuses or stays silent?

We push back with the dated evidence and the legal ground, report again under a second rule where one applies, and go to the company that hosts the content. If a covered platform misses the 48-hour deadline on an intimate image, we report it to the FTC at TakeItDown.ftc.gov. Every step lands in your evidence file. The 48-hour rule, step by step.

Do I have to send you the video or the image?

No. Never send it to us. Send the links to the pages where it appears, and we take dated screenshots ourselves. If you have the file, StopNCII.org can create a fingerprint of it on your own device so partner platforms block copies. We walk you through that too.

Someone used my face or voice in a scam ad. Can you help?

Yes. A fake ad or account that uses your face, your voice or your name to sell something is impersonation. We report the ad, the account and the landing page, and we tell you what to say to the people who got fooled. Same price, same seven days.

A fake account or store is copying my small business. Is this for me?

Yes. We report the fake account to the platform, and the fake store to its host and to Google. If you want daily monitoring after the seven days, the Business Shield exists. We mention it once, in your day-7 report, never before.

Is there a free way to do this?

Yes, and we will tell you about it. StopNCII.org creates a fingerprint of an intimate image on your device so partner platforms can block copies. Loti AI has a free tier that searches for your face and helps you file reports yourself, and Alecto AI is a free assistant that helps you prepare takedown requests. Every platform's own report form is free. What you pay us for is a dedicated handler who files, follows up for seven days and keeps the proof, without ever asking for your face.

If you need help right now

Immediate danger: 911. Crisis or suicidal thoughts: call or text 988. Sextortion or blackmail: do not pay, report at ic3.gov. Anyone under 18 in the image: takeitdown.ncmec.org and report.cybertip.org, free. Someone to talk to about intimate images shared without consent: Cyber Civil Rights Initiative helpline, 1-844-878-2274. We do not take blackmail or under-18 cases, and we do not charge for telling you so.

Want to do it yourself first? The first hour: a free one-page checklist (PDF), and what to do when the platform ignores your request.

A deepfake of you is online. Start this hour.

Send us the links. Open a case, get a dedicated handler within the hour, 8 am to 10 pm Eastern Time. First reports within two hours. $299, one time, charged in US dollars.