§ Free check · 30 seconds
Did AI steal your book?
We cross-reference your book against the datasets Meta, Anthropic, and others admitted usingto train their AI models. If your book’s in there, you may be eligible for active class-action lawsuits — the Anthropic settlement paid ~$3,000 per claimed title.
13 datasets
Cross-referenced
Real Books3 match
197k titles, ISBN + filename
Letter templates
Pre-filled, ready to send
About this check
- What we check: 13 datasets including LibGen, Z-Library, Books3, BookCorpus, HathiTrust, Common Crawl, Hugging Face, AO3, Wattpad, Internet Archive scans, and the Anthropic settlement class.
- How we check:for Books3 (used by LLaMA-1, Bloomberg GPT, EleutherAI), we directly search 197,000 filenames + 131,000 ISBNs we’ve indexed. Other datasets are heuristic (publication year, ISBN, Amazon listing) while we complete the full Anna’s Archive ingest.
- What we DON’T claim:we can’t prove your book was used to train any specific model — nobody can without the model weights. We can prove your book is in the datasets the AI companies admitted using. That’s the same evidentiary basis the Authors Guild used in the Anthropic case.
- Your email: stored so we can send the check result + occasional tips on AI discoverability for authors. Unsubscribe anytime; we never sell email lists.
Want the full AI Discovery audit?
The scrape check is one chapter. The full audit also tells you whether AI engines recommend your book to readers — across ChatGPT, Claude, Perplexity, and Gemini. £29.99, one price, no upsells.
See the full audit →