Every quarter Privly publishes a snapshot of the leak-site ecosystem from data sourced directly from Google's Transparency Report. This is the Q2 2026 edition — the first public report. Together, the 50 sites we track have had over 77 million URLs requested for delisting from Google search. 4.9 million of those requests came in the last four weeks alone. The trends inside that data tell a different story than the conventional wisdom about creator-content piracy. This report walks through what the numbers actually say.
If you want to interact with the underlying data — sort it, filter by category, search by domain — head to /observatory. This article is the editorialised version: the patterns, the surprises, the implications for creators.
The headline numbers
77,071,655 URLs requested for delisting from Google Search across our 50-site cohort, all-time. 4,906,829 URLs requested in the last four weeks. That's roughly 1.2 million new takedown URLs every week against this small slice of the leak ecosystem. The cohort spans aggregators, tube sites, file hosts, forums, and deepfake-specific sites. Of the 50 sites, 45 had active Google data; five returned no records at all (mostly newer or smaller domains where no creator has ever filed a DMCA).
These numbers come from a public source: each per-domain page on transparencyreport.google.com/copyright/domains/. Anyone can verify any individual site's count. We pulled the data, aggregated it, and now publish the synthesis.
Finding 1: aggregators dominate, not tube sites
The biggest surprise in the data is how completely the leak aggregators outweigh the tube sites. Conventional creator-protection writing focuses on the household names — Pornhub, xHamster, xVideos, RedTube, YouPorn, Tube8 — as if those are where leaked content lives. They're not.
Our cohort includes eight Aylo-owned and adjacent tube sites. Combined, all eight have around 10 million all-time URL takedowns. The aggregator category — Fapello, Coomer, Kemono, Erome, Bunkr, Thothub, Nudostar, Wildskirts, Leakedzone, Fapodrop, and similar — has over 65 million all-time URL takedowns. That's roughly 6× the volume of the tube-site category, in a category with fewer sites.
Fapello.com alone has 24,571,887 URL takedowns — more than every tube site on our list combined and multiplied by about 50. Nudostar.tv (13.7 million), Coomer.su (6.2 million), Wildskirts.com (4.7 million), Leakedzone.com (3.7 million), and Fapodrop.com (3.4 million) round out the rest of the leaderboard. None of these are household names. All of them are bigger creator-content distribution problems than the tube sites people warn about.
The implication for creators is direct: when you're triaging where to focus DMCA enforcement effort, the household-name tube sites are the easy targets — they have compliant DMCA agents and respond to notices reliably. The aggregators are harder, more numerous, and where most of your content actually ends up. Most creators get this backwards.
Finding 2: recent activity tells a different story than all-time totals
The lifetime takedown count is a noisy signal. It mixes content that's been actively leaked for ten years with content that arrived last week. The "last four weeks" column tells you which sites are active threats today.
The data shows a sharp bimodal pattern. Some sites that look big on lifetime data are essentially dormant. Bunkr.ru has 24,536 lifetime takedowns but only 10 in the last four weeks — Google has stopped seeing meaningful new content on this specific domain (most likely because the operator moved to a mirror). Thothub.tv: 16,397 lifetime, 1 in the last four weeks. Cyberdrop.me: 14,361 lifetime, 2 in the last four weeks. Mrdeepfakes.com: 33,382 lifetime, 20 in the last four weeks. These sites still rank high on the all-time list but aren't where new content is landing today.
Other sites are getting hammered. Fapello.com had 2.35 million takedowns in the last four weeks — roughly 587,000 per week. Nudostar.tv had 1.97 million in four weeks. Eporner.com (210K), Leakedzone.com (99K), Fapodrop.com (75K), Erome.com (39K), Nudostar.com (35K), Coomer.st (30K), Wildskirts.com (30K), and Spankbang.com (11K) round out the most-active sites. If you're a creator allocating DMCA enforcement effort right now, these are where the attention belongs.
The "is this site actively dangerous to creators today?" question is answered in the recent four-week column, not the lifetime total.
Finding 3: mirror domains fragment DMCA volume
This is the most operationally important pattern in the data, and it's invisible if you only look at one domain at a time.
Operators register multiple TLD variants of the same site so DMCA filings split across them. We tracked eight pairs that are clearly the same operator running parallel domains:
- bunkr.ru (24,536) vs bunkrr.org (1) — virtually all volume on one TLD - coomer.su (6.2M) vs coomer.st (974K) — primary still dominant but mirror is real - kemono.su (475K) vs kemono.party (350K) — roughly even split - simpcity.cr (8.7K) vs simpcity.su (414K) — vast majority on the .su, opposite of canonical naming - thothub.tv (16K) vs thothub.lol (911K) — the .lol mirror is 50× bigger than the listed primary - fapello.com (24.6M) vs fapello.su (1.1M) — primary dominant - cyberdrop.me (14K) vs cyberdrop.to (6K) — similar scale - thotsbay.com (40K) vs thotsbay.tv (1.8M) — mirror is 45× bigger
Filing only on the canonical domain misses content on the mirror. In several cases the "mirror" actually has more volume than the original. The operational implication is that your DMCA workflow has to enumerate every active TLD variant of an operator and file against each one separately. Anything else leaves content up.
Removing a URL from bunkr.ru while leaving the same content on bunkrr.org accomplishes nothing — but a service that only files on the primary will report it as a successful removal.
This is also why aggregate "removal success rate" numbers from protection services need scrutiny. Removing a URL from bunkr.ru while leaving the same content on bunkrr.org accomplishes nothing — but a service that only files on the primary will report it as a successful removal.
Finding 4: tube sites have a long tail, aggregators have a peak
A subtle but useful pattern: tube sites have steady, slow DMCA volume going back to 2011 (770 weeks of data). Aggregators have shorter histories — most appeared between 2018 and 2022 — but their per-week volume is 10-100× higher when active.
Tube sites: established, regulated, slow-moving. They're not going anywhere, but they're not where the action is. Aggregators: faster, hotter, younger. The threat profile creators face today comes from sites that didn't exist five years ago and may not exist five years from now — but while they're operating, they're far more dangerous per-week than anything Aylo runs.
This shifts how a sustainable creator-protection strategy should be designed. You can build a once-and-done DMCA workflow against tube sites that keeps working for years. Aggregators require continuous adaptation — new mirrors, new TLDs, new sites. The reactive model that worked for tube-site enforcement doesn't work for aggregators.
Finding 5: the SEO gap in our own ecosystem coverage
Privly currently has dedicated /remove-from-* pages for 7 leak sites: bunkr, coomer, cyberdrop, erome, simpcity, thothub, mrdeepfakes (plus Telegram and Google as platforms). Of the top 10 sites by all-time takedown volume, only 2 are covered by these dedicated guides (coomer.su and erome.com). The other 8 — fapello.com, nudostar.tv, spankbang.com, wildskirts.com, leakedzone.com, fapodrop.com, eporner.com, thotsbay.tv — represent millions of takedown requests and clear creator search demand for "how to remove [my content] from [site]" queries. Building those eight guides is one of the highest-leverage SEO investments we can make over the next quarter.
This is the kind of finding that's hard to see without aggregating the data. Looking at any one site doesn't tell you which guide is missing. Looking at all 50 ranked, the gap is obvious.
What the data doesn't tell us
A few honest limits worth being explicit about.
These are takedown requests, not confirmed removals. Google's per-request delisting rate varies — some requests get fully actioned, some get rejected, some get partial action. The headline number is "URLs creators and rightsholders have asked Google to remove" — that's what's being measured, no more, no less.
Google groups data by exact domain. Subdomains and TLD variants are separate datasets. We've highlighted the mirror-domain pattern explicitly because it matters; readers should understand the same applies to any operator who shows up under multiple TLDs.
The data doesn't tell us about non-Google takedowns. DMCA notices sent directly to a leak site, to its host, or to its CDN aren't reflected in the Google numbers. Most of the actual takedown work creators and protection services do goes to non-Google channels. Our internal Privly data shows us those numbers; this report is specifically about the Google-visible slice.
The data doesn't capture the Telegram channel ecosystem, Discord server distribution, or non-Google search visibility. All three are important parts of the creator-leak picture and not tracked by the Transparency Report.
What we're going to do with this data
Three commitments based on this report.
First, we'll build the missing 8 /remove-from-* pages over the next 60 days, starting with fapello (the single largest target by far). Each page becomes a guide for creators searching for help removing content from that specific site, plus a landing page that internal-links into the broader Privly content network.
Second, we'll keep republishing this report quarterly. Q3 2026 will land in roughly 90 days. Year-over-year comparisons start mattering in Q2 2027, when we have a year of trend data on the same 50 sites.
Third, we'll add new domains to the cohort as the ecosystem shifts. If a new aggregator emerges in Q3 with significant takedown volume, it goes in the next quarter's report. If existing sites get shut down or demonstrably dormant for two consecutive quarters, they'll be flagged in the report and eventually removed from active tracking.
Sourcing and replication
If you want to verify any number in this report or replicate the methodology, the source data lives at transparencyreport.google.com/copyright/domains/[domain]/ for any of the 50 sites in our cohort. Our methodology is documented at /observatory under "Methodology." The full machine-readable dataset is publicly accessible and free to use under Creative Commons Attribution 4.0 — cite "Privly Observatory" and link back when you do.
For journalists, researchers, and other content-protection services: this data is yours to use. We built it because the conversation around creator content protection deserves better numbers than the ones competitors quote without sourcing. If you find an angle, an error, or a question we didn't address, get in touch.
