In May, thirteen major publishers won a massive $19.5 million default judgment against shadow library Anna’s Archive in a New York federal court.

This week, the same publishers, including Penguin Random House, Elsevier, and HarperCollins, filed a new complaint at the same court, this time with the relatively young pirate library WeLib as the target.

Again, the stakes are substantial, with the publishers seeking up to $19.5 million in potential damages for direct copyright infringement.

A New Entrant

The similarities don’t stop at the legal arguments and stakes. Anna’s Archive already highlighted the newcomer in a blog post last year, describing WeLib as a “new entrant” in the space that had copied both its collection and its code.

“They appear to have mirrored most of our collection, and use a fork of our codebase,” Anna’s Archive noted.

The same blog post was also critical of WeLib for not contributing back to the ecosystem and recommended that people avoid using the site.

This week, the publishers also warn against using the site, albeit for different reasons. Their complaint accuses WeLib’s unnamed and anonymous operators of widespread copyright infringement, while also confirming that connection to Anna’s Archive.

“Defendants’ entire business is the illegal copying and distribution of literary works,” the complaint notes, adding that “WeLib was created after its operators copied the source code and most of the contents of the Notorious Pirate Site, Anna’s Archive.”

Not a Library

WeLib describes itself as an “endless library” founded on the principle that “education and literature belong to everyone.” The publishers, however, clearly don’t agree with the library framing, noting libraries can be trusted; pirate sites not.

“Libraries are trusted institutions that serve the communities that fund them by lending books and other publications they have lawfully acquired. Using this label for WeLib explicitly misleads the public and allows WeLib to hijack the goodwill that libraries enjoy and have legitimately earned.”

“WeLib is no more than a pirate website that reproduces and distributes works of authorship owned by others to users for a profit, without authorization from or compensation to the copyright owners,” the complaint adds.

The complaint notes that WeLib’s operators made efforts to keep their identities hidden. However, the site itself quickly became a go-to portal for many book pirates.

The complaint notes that, by WeLib’s own account, its collection includes 43 million books and 98 million articles. The site reportedly has over 80,000 active monthly users who accessed more than 51.7 million books and downloaded 14.5 million files last month.

While the site can be used for free, users can pay for fast downloads and to skip the queue. Subscriptions start at $7 per month for 25 fast downloads and 25 fast reads per day; while the top tier costs $90 a month for 1,000 daily downloads.

These payments, or “donations” as WeLib calls them, can be made through cryptocurrency, WeChat, and Alipay. They are allegedly processed through a company called Malum.co, which offers payment services to high-risk vendors, without the need for any KYC identity checks.

Damages and Domain Seizures

The complaint lists a sample of 130 copyrighted works as evidence. This mirrors the Anna’s Archive lawsuit, where the court awarded $150,000 per work, which is the statutory maximum, resulting in a total of $19.5 million.

In addition to the monetary damages, the publishers are also seeking a permanent injunction that aims to take the site offline. They ask the court to order third-party registries, registrars, and hosting providers to disable WeLib’s domains and render them untransferable.

This also includes a specific request to disable the authoritative nameserver for the .st domain, registered through Njalla, a Costa Rica-based registrar that is not necessarily responsive to U.S. court orders.

The AI Training Conundrum

As with other recent publisher lawsuits, the complaint also mentions AI training. Specifically, it alleges that WeLib supplies copyright infringing data to AI companies.

“WeLib has also been an illegal supplier of stolen content to the AI industry. In a recent lawsuit, publishers alleged that Meta utilized WeLib to train their Llama models,” the complaint reads.

The recent lawsuit they refer to is Elsevier Inc. v. Meta Platforms which is filed by several of the same publishers through the same law firm, Oppenheim + Zebrak. However, what that complaint actually says about WeLib is more specific and not in line with the current case.

The Elsevier v. Meta complaint describes WeLib as a source found within C4 training dataset Meta used, but identifies it as “formerly known as PDF Drive.” This dataset was built years ago from a Common Crawl snapshot and predates WeLib and even Anna’s Archive.

More confusingly, the complaint against WeLib that was filed this week makes no mention of it formerly being known as “PDF Drive”, or the C4 dataset for that matter.

According to our knowledge, there is no evidence that content hosted by WeLib was included in the C4 database. All we can confirm is that the database does include “PDF Drive” data and that the pdfdrive.com domain redirected to the new WeLib site at some point.

PDF Drive is a long-running PDF hosting site that has operated for years, predating Anna’s Archive entirely. It has no documented connection to Anna’s Archive’s codebase or collection. Whether it shares more than a domain redirect with the WeLib now being sued is unclear.

The publishers’ framing of WeLib as an active AI training pipeline may be getting ahead of the evidence. For now, WeLib has yet to respond. However, since anonymous operators typically don’t show up in court, this case may also copy Anna’s Archive’s path, heading to a default judgment.

—
A copy of the complaint, filed by Oppenheim + Zebrak on behalf of the thirteen plaintiff publishers, is available here (pdf) .

From: TF , for the latest news on copyright battles, piracy and more.