Deepfake removal, and why a DMCA notice is usually the wrong instrument

    None of the six creator-protection services compared here publishes a deepfake detection or removal product, and that includes openDMCA. A deepfake using your face is often not a copyright matter at all, because you may not own the underlying video, so the working route is usually the platform's non-consensual imagery process.

    What six creator-protection services publish about detection capability, alongside price, free trials, published surfaces and review evidence
    FeatureopenDMCARultaBranditScanBruqiCeartasPrivly
    Entry price$29/mo — Starter, 1 creator$139/mo — Pro, 1 username$69/mo — Premium, 3 stagenames$29/mo — Side Hustle$69/mo — Star, 1 creator$49/mo — Creator Protection
    Price to cover three stage names$89/mo — Pro, up to 3 creators$324/mo — Legend, up to 4$69/mo — Premium covers 3$29/mo — all usernames includedNot published$49/mo — all platforms, all content
    How the price scalesFlat plan, then $15/mo per extra creatorPer usernameBy stagename countFlat; scan frequency is the upgradePer creator, per monthOne flat fee, no per-username surcharge
    Something you can operate before payingFree username scan, no accountFree leak search before signupFree leak check, plus a 7-day trial stated to need no cardFree DIY tier, no expiryTrial referenced, terms not published7-day trial, stated to need no card
    Published Telegram routeRouted to Telegram's published DMCA address"Telegram Takedown" — Legend tier onlyNot publishedNot publishedNot publishedNot published
    Face or image matching, as a published featurePerceptual hashing of images you choose, plus reverse image search"Reverse Image Search" — Legend tier only"Facial Recognition & AI Detection" — PremiumNot publishedNot publishedNot published
    Impersonation or fake-account removal, as a published featurePublished guide, not automated"Social Media Removals" — Premier tier and above"Social Media Removals" (Instagram, X, Reddit) — Premium"Impersonators" — Creator Pro and aboveNot publishedNot published
    Published price for a ten-creator roster$229/mo, 10 creatorsNot published — booked callNot publishedNot publishedNot publishedNot published
    Independent public review profileNoTrustpilot 4.6 from 140 reviewsTrustpilot 4.6 from 68 reviewsTrustpilot 3.7 from 48 reviewsTrustpilot 4.3 from 27 reviewsProfile exists, 0 reviews

    Choose someone other than openDMCA, because openDMCA does not sell this if…

    • You need synthetic-media detection as a product. openDMCA does not have one. AI-likelihood scoring exists in its codebase but depends on a worker service that is not enabled in production, so it is not offered, not sold, and not being described here as a capability.
    • Face matching is close enough to what you need. BranditScan names "Facial Recognition & AI Detection" on its $69 plan and is the nearest published thing in this comparison. Ask what it actually detects before you buy on the phrase.
    • The material is intimate content of an adult, made or shared without consent. The correct first route is the platform's own non-consensual imagery process, not a takedown vendor, and it is free. Report there first and treat any paid service as a supplement.
    • The person depicted is a minor. Stop reading comparison pages. Report to the relevant national reporting centre immediately; that is the only route that matters and no commercial service substitutes for it.

    BranditScan

    Best for: Names AI detection, but not deepfakes. "Facial Recognition & AI Detection" is listed on the $69 Premium tier. What it detects is not described, and detecting a face is not the same as identifying synthetic media.

    Pricing reality: Premium: $69/mo for 3 stagenames. Premium (annual): $690/yr for 3 stagenames. White Glove: $149/mo for Unlimited stagenames. White Glove (annual): $1490/yr for Unlimited stagenames.

    What it does better: Three stage names are included in the $69/mo entry plan, where a per-username competitor would charge for each.

    What makes people leave: On deepfakes specifically: BranditScan publishes no detection or removal product for synthetic media on the pages checked here, so anything you are told verbally about deepfake coverage is not something you can hold them to. branditscan.com on BranditScan

    Verdict: No published deepfake product. Judge on the rest of the sheet and use the platform routes below for the synthetic media itself.

    Rulta

    Best for: Impersonation and image search, no deepfake product. Rulta publishes reverse image search on Legend and runs a catfishing and impersonation service. Neither is a deepfake product and Rulta does not describe one.

    Pricing reality: Pro: $139/mo for 1 username. Premier: $184/mo for 2 usernames. Legend: $324/mo for 4 usernames.

    What it does better: Rulta states it is "Google's Trusted Copyright Partner" and cites Google's public Transparency Report for a claimed 200,000,000+ links reported.

    What makes people leave: On deepfakes specifically: Rulta publishes no detection or removal product for synthetic media on the pages checked here, so anything you are told verbally about deepfake coverage is not something you can hold them to. rulta.com on Rulta

    Verdict: No published deepfake product. Judge on the rest of the sheet and use the platform routes below for the synthetic media itself.

    Bruqi

    Best for: Impersonator coverage from $99, nothing on synthetic media. Creator Pro names impersonators among its surfaces. The pricing page says nothing about deepfakes or AI-generated content at any tier.

    Pricing reality: DIY: $0/mo for Manual removals only. Side Hustle: $29/mo for All usernames. Creator Pro: $99/mo for All usernames. Top 1%: $249/mo for All usernames.

    What it does better: Every paid tier covers all usernames at one price, starting at $29/mo — the flattest pricing model of the five competitors compared here.

    What makes people leave: On deepfakes specifically: Bruqi publishes no detection or removal product for synthetic media on the pages checked here, so anything you are told verbally about deepfake coverage is not something you can hold them to. bruqi.com on Bruqi

    Verdict: No published deepfake product. Judge on the rest of the sheet and use the platform routes below for the synthetic media itself.

    Ceartas

    Best for: Nothing published. The pricing page publishes tiers and manual-request caps and describes no detection capability of any kind.

    Pricing reality: Star: $69/mo for 1 creator. Elite: $169/mo for 1 creator. VIP: $349/mo for 1 creator. Platinum: $1200/mo for 1 creator.

    What it does better: A published price ladder that reaches genuinely high-touch service, which most of the category does not publish at all.

    What makes people leave: On deepfakes specifically: Ceartas publishes no detection or removal product for synthetic media on the pages checked here, so anything you are told verbally about deepfake coverage is not something you can hold them to. ceartas.io on Ceartas

    Verdict: No published deepfake product. Judge on the rest of the sheet and use the platform routes below for the synthetic media itself.

    Privly

    Best for: Nothing published. The flat $49 plan states all platforms and all content, which is not a claim about synthetic media in either direction.

    Pricing reality: Free Trial: $0/mo for 7 days, 1 scan, up to 3 uploads. Creator Protection: $49/mo for All platforms, all content. Platform Offboarding: $499 one-time for Concierge, 12 months of cover. Copyright Registration: $299 one-time for Optional add-on.

    What it does better: The flattest published price in this comparison for a full-service plan: $49/mo, in Privly's words "Same plan, every feature, no per-username surcharge, no hidden tiers".

    What makes people leave: On deepfakes specifically: Privly publishes no detection or removal product for synthetic media on the pages checked here, so anything you are told verbally about deepfake coverage is not something you can hold them to. useprivly.com on Privly

    Verdict: No published deepfake product. Judge on the rest of the sheet and use the platform routes below for the synthetic media itself.

    The objections worth taking seriously

    Is this a scam, like half the category?
    Fair question, and the honest answer is that openDMCA cannot prove it with a testimonial. What it can do is publish the instrument: a notice under 17 U.S.C. §512 is the same one studios file, openDMCA acts as your authorised agent, and the policy that binds openDMCA to that role is public. Read the openDMCA DMCA policy, including the counter-notice route
    Will filing put my legal name in front of the people who leaked me?
    A valid notice has to identify a complainant, and openDMCA is the named agent on it rather than your home address. Nothing about your identity handling is left to a promise — the security page describes what openDMCA stores about you and where it goes. How openDMCA handles creator identity and account data
    Do I have to upload my leaked content to your servers?
    No. A openDMCA scan takes a username, not a file. Reference images are optional and only used if you choose face or image matching, and the privacy policy states plainly what is stored and for how long. What openDMCA stores, and what it never asks you for

    The routes that actually exist, in the order to try them

    None of this is legal advice, and deepfake removal can involve rights that a takedown service cannot act on at all. What follows is the practical sequence, and the first two steps are free.

    First, work out which instrument applies. A copyright notice removes material you own the copyright in. If a deepfake was built by pasting your face onto someone else's video, you may not own the underlying footage, and a DMCA notice may have no basis. If it used your own published photos or video as source material, a copyright route may exist for those elements. The free DMCA or NCII decision tree walks that fork without asking for anything.

    Second, use the platform's non-consensual imagery process. Every major platform publishes one, it is separate from copyright, and it does not depend on who owns the footage — which is exactly why it works in cases where a notice does not. It is also free and usually faster than a commercial route.

    Third, go after the search path. Even where the host will not act, search de-listing is a separate route with its own outcome. Most people find a deepfake through a search result rather than by browsing the site hosting it, so removing the search path cuts reach materially even while the file remains. The free de-listing request builder produces the submission text.

    Fourth, document before anything comes down. URLs, timestamps and screenshots, captured before a removal succeeds. Once material is gone it is much harder to evidence that it existed, and evidence is what any later step depends on.

    What to ask a vendor selling deepfake coverage. What does the detection match on. Does the workflow use the copyright route or the non-consensual imagery route, and who decides which. What happens when a host refuses. A vendor who answers all three has thought about it; a vendor who answers with a feature name has not.

    What openDMCA does and does not do here, stated plainly

    openDMCA does not offer deepfake detection. The AI-likelihood scoring in its codebase depends on a worker service that is not enabled in production, which makes it a capability openDMCA has written and does not run. Selling it, or describing it as available, would be the exact thing this whole comparison layer exists to avoid.

    What openDMCA does do is route. openDMCA routes every infringing URL through the Removal Graph — a live map of which hosts actually remove content, how fast, and through which channel — so each takedown takes the shortest path that has historically worked for that host. If a synthetic image of you sits on a host with a published abuse route, that URL is handled like any other: host notice, escalation, provider or CDN, then search de-listing, with a host below a 25% removal rate over ninety days across at least five resolved samples fast-pathed to the last rung. Perceptual hashing against reference images you choose to provide will find re-uploads of a specific image once you have one.

    That is a partial answer to a deepfake problem and it should be read as one. The detection question — is this synthetic, and is it me — openDMCA does not answer today.

    Why this page has no recommendation. Every service in the comparison table publishes prices, surfaces and review histories, and none of them publishes a deepfake product. Ranking them on a capability nobody advertises would be inventing a distinction to fill a page. The comparison table is here because the rest of the sheet still matters if you are buying anyway, and the routes above are here because they are what actually works.

    The estimate below assumes 60 live copies, a smaller count than a leaked-set page because synthetic material typically propagates across fewer hosts before search picks it up.

    Left up, 60 exposed copies work out to an estimated $351 a month in diverted subscriptions. Estimate, not a measurement: it assumes 0.45 diverted subscribers per exposed post at an average subscription price of $12.99. Your number depends on your pricing and audience.

    Frequently asked questions

    Which service removes deepfakes?

    None of the six compared here publishes a deepfake detection or removal product, and that includes openDMCA. BranditScan names "Facial Recognition & AI Detection" on its $69 plan, which is the closest published capability, but detecting a face is not identifying synthetic media.

    Can I file a DMCA notice against a deepfake of me?

    Only if you own copyright in the material used. If your face was pasted onto footage you do not own, a copyright notice may have no basis. Where your own published photos were the source, a route may exist for those. This is not legal advice.

    What works if copyright does not apply?

    The platform's own non-consensual imagery process, which is separate from copyright and does not depend on who owns the footage. It is free, it is usually faster than a commercial route, and it is the correct first step for intimate synthetic material.

    Is search de-listing worth pursuing for a deepfake?

    Usually yes. Most people encounter this material through a search result rather than by browsing the host, so removing the search path cuts reach substantially even where the host will not act. It is a separate route with its own outcome.

    Should I screenshot before reporting?

    Yes. Capture URLs, timestamps and screenshots before anything comes down. Once material is removed it becomes much harder to evidence that it existed, and every later step — a platform appeal, a legal route — depends on that record.

    Why is openDMCA not selling me something on this page?

    Because it does not have the product. The detection code exists but depends on a service that is not enabled in production, so claiming it would be a capability openDMCA cannot deliver. The routes above are free and they are what actually works today.

    Sources and verification dates

    Report through the platform's non-consensual imagery process first — it is free and it does not depend on who owns the footage. For the leak-removal half of the problem, Starter is $29/mo and Pro $89/mo.