What a match means

If your name turns up in a LibGen catalog snapshot, the PiLiMi index or the Books3 file list, you will want to know what that says about your work. A match is a useful, concrete lead. It is also narrower than it may first appear.

Three separate questions

1. Was the work listed? A title, author, ISBN or file name matches an entry in a dated catalog or file list. This is the question a PayAuthors report answers.
2. Did a particular company copy it? That needs evidence tying the company to the collection, such as its own papers, testimony or a court record.
3. Was the copying unlawful? That depends on ownership, authorization and defenses such as fair use, and is a question for a lawyer.

What a match shows

A LibGen line means the Library Genesis non-fiction or fiction catalog listed a file with that title and author, and the line names the dated snapshots in which it appears (we hold 25, from November 14, 2015 to June 3, 2025), so you can see when it was first listed and whether it was listed in the snapshot nearest a documented download (the timeline shows how near each snapshot is). A PiLiMi line means the record is in the Pirate Library Mirror index of Z-Library, and says whether it shipped in the July 2022 release or the September 2022 release. A Z-Library line means the record was added to Z-Library after PiLiMi: it is in Anna’s Archive’s index of records added from August 2022 to August 2023 (with the date added), or in the January 2025 Z-Library catalog as published in a research catalogue. An Internet Archive line means the Archive’s lending library lists a scanned copy of the book, with the date it was added; the Archive is a nonprofit library, and a listing there is not a finding that anyone did anything wrong. A Books3 line means a file whose name carries the name you searched is in the published list of Books3 file names. A registration line is a candidate match to a public U.S. Copyright Office record (registrations 1978 through 2025), with a link so you can check it, and a timing mark when the registration date is early enough for the statutory-damages rule against a documented download date.

What a match does not show

A match does not show that any company downloaded the file, that the book was used to train an AI model, or that anyone infringed your copyright. A registration match is not a finding of authorship or ownership, and a missing match does not mean a book is unregistered or absent from other datasets. The report also lists every record matching the name as typed, so a common name can bring up other people’s books.

It is not perfect. Matching is automated: expect some missed records and some matches that are not yours, and check anything you rely on.

Timeline of the AI copyright cases

Each documented event on its own row, with how precisely its date is known; our dated catalog snapshots along the bottom. A dashed line connects an event to the latest snapshot of that source before it. A paid report adds, in each snapshot box, the number of candidate records under your name, and reads the comparison for you.

Why it still matters

Where a company’s own statements or a court record show it copied an entire collection, one remaining question for an author is narrower: was my work in that collection? Ownership, registration and defenses such as fair use still have to be addressed. A dated, documented match helps answer exactly that. The court record summary in our Methodology explains how courts have treated these allegations.

Registration before suit. Separately, a U.S. work generally can’t be sued on until the Copyright Office registers or refuses it (17 U.S.C. § 411(a)). Works first published outside the United States are exempt from that requirement but not from § 412, so they can still claim actual damages and are not in our “works” counts. Our registration data runs through 2025.

Left out of the Anthropic settlement?

The class in Bartz v. Anthropic (N.D. Cal.) was defined narrowly. To be on the settlement's Works List, a book had to be in the LibGen or PiLiMi copies Anthropic downloaded, carry an ISBN or ASIN, and be registered with the U.S. Copyright Office within five years of publication and either before the settlement’s Download Date (August 10, 2022, a date the agreement sets for every book) or within three months of publication. Roughly seven million books were downloaded; the final list covered 482,460 works. Books were left off for reasons that had nothing to do with whether they were copied: no ISBN, a foreign edition, a registration that came too late, a work registered only as part of an anthology or other collective registration, or a registration and catalog entry that did not match up. Authors whose books did not make the list kept their claims, and some are pursuing them individually or in groups, as are authors who opted out.

A PayAuthors report shows every LibGen catalog entry, PiLiMi record and Books3 file under a name, ISBN or not, with the dates each was listed and candidate registration matches to check against the Copyright Office record. If your book was in LibGen in June 2021 or in the July 2022 PiLiMi release but not on the Works List, the report is a starting point for that conversation with a lawyer. Sources: class certification order (July 17, 2025), settlement notice and final approval order (July 20, 2026); the Authors Guild's settlement FAQ explains the common reasons for exclusion.

If you find a match

Keep the report. Compare each line with your own publication and registration records: title, edition, ISBN, publisher and registration number. Check candidate registrations against the linked Copyright Office record. If you are considering a claim or facing a deadline, talk to a copyright lawyer.

This page is general information, not legal advice.

Check a name for free