Products
Our proprietary Croquis tool detects creative works inside AI training datasets and models. We search the giant public datasets used to train AI, and we probe the models themselves—across modalities, domains, and scales. If work has been taken, we can show it, with evidence.
Protection isn’t one-size-fits-all. The same technology works for a single artist watching over their own portfolio, a rights organization managing thousands of works, and a legal team that needs findings solid enough to build a case on.
Protect your work and your style. See where your work has appeared, get the report, and decide what happens next.
See Individual creatorsStudios, publishers, estates, and rights organizations: audits and continuous monitoring across catalogs of thousands of works.
See EnterpriseLaw firms and litigation teams: screen potential claims and build documented, reproducible evidence for the works at issue.
See LegalWe look for names and works in public training datasets and probe AI models directly using membership inference—the methods we explain in plain language in our FAQ.
Transparent, per-work reports with supporting evidence: which datasets and models we checked, what we found, and how confident we are.
As new models and datasets appear, the analysis runs again. And for work published on your own sites, we add safeguards at the source to resist scraping.
A single portfolio, a catalog of thousands of works, or the works at the center of a claim—across text, image, and other modalities.
Croquis checks the public datasets used to train AI models and probes the models themselves, and keeps checking as new ones appear.
Use the reports to guard work going forward, set licensing terms from a position of knowledge, or take the findings to your legal team.
Protection starts with visibility—whoever you are and whatever scale you work at. See a sample of our findings here, or start the conversation.
