The New York Times, Daily News, Chicago Tribune, and Orlando Sentinel filed the complaint in Manhattan federal court. They describe Brave Software as a supplier, dealer, and distributor of stolen, copyrighted content.

The filing alleges Brave built a crawling index of more than 30 billion websites and sold access to that index to Anthropic, Perplexity, Mistral, xAI, and Cohere. The New York Times has previously sued OpenAI and Perplexity on similar grounds.

The complaint also targets Brave browser features that let users read paywalled articles without paying.

What the filing does not establish: no proof any of the five labs actually purchased the index, no price, no volume, no contract term. The document leaves open whether Brave's crawler technically bypassed paywalls or merely indexed publicly accessible pages that the publishers later gated.

No docket number, no damages figure, no count of pages crawled from the four plaintiffs, and no Brave revenue line appear in the filing as reported.

The legal risk for the named labs sits upstream of model weights. If the court accepts Brave as a content intermediary, each recipient inherits a supply-chain question the complaint leaves open: whether the training data was licensed, scraped, or both.

The draft repeated two paragraphs nearly verbatim and buried the most quotable line — the 'supplier, dealer and distributor' language — in the middle of a dense first paragraph. The rewrite leads with that accusation, separates what the complaint alleges from what it cannot prove, and keeps the paywall angle as a distinct beat rather than a clause tacked onto the crawling allegation.