Executive Summary

Key Figures

Source: reporting on the two complaints filed on November 2, 2026, by TNW, Search Engine Journal and CyberInsider

Two complaints naming Brave Software as defendant reached Manhattan federal court on November 2, 2026. One is a 72-page filing from The New York Times alone, docketed as 26-cv-9549. The other is a 69-page filing brought jointly by the New York Daily News, the Chicago Tribune and the Orlando Sentinel, docketed as 26-cv-9554; all three of those papers belong to Alden Global Capital. There are two cases, but they point at the same company and lay out nearly the same facts.

The language the complaints use for Brave is not polite. The plaintiffs call the company a supplier, seller and distributor of stolen works, and reach for the image of a shop that moves stolen goods, labeling Brave an AI fence. A Times spokesperson said the paper supports the ethical and responsible use and development of AI, then said Brave had disregarded current copyright law and had not responded to the cease-and-desist requests the paper sent more than once. The complaint dates those requests to August 2025 and April 2026.

The complaints lead with the size of the index: more than 30 billion web pages, of which over 2 million are articles from the three Alden papers. The crawler that filled the index has never been publicly identified, the complaints say. That point is more than a grievance, because a publisher can block a particular crawler only if the crawler announces its own name. Against an unnamed collector there is no line to write in robots.txt.

Brave's answer on this point is already public. In its help documentation the company explains why its crawler carries no distinct user agent: a fair number of sites open their doors only to Google's crawler, and a separately labeled bot would be turned away in those places. Instead, Brave writes, its bot obeys whatever robots.txt rules apply to Googlebot. Read from the publisher's side, that answer means blocking Brave requires blocking Google too, which is not a road anyone who depends on search traffic can take.

A browser company holding a 30-billion-page web index may be the strangest part of the story. Five years of history stand behind it. In March 2021 Brave acquired Tailcat, formerly Cliqz, a German team that had been building an independent search engine. Brave Search opened that June, leaning on outside engines such as Bing for roughly 13 percent of its queries, and that share fell to 7 percent in 2022.

The turning point came in April 2023. Microsoft, newly partnered with OpenAI, raised the price of the Bing search API sharply, and Brave moved its text results to 100 percent of its own index. Images and video followed that August. Once the company stopped renting someone else's index, the next move was to sell its own. The Brave Search API launched in May 2023.

That API now reaches deep into generative AI. Alongside the endpoints that return web and news results, Brave sells LLM Context, which pulls out the passages that match a query and scores their relevance, and Answers, which hands back a finished AI answer with sources attached. This is the grounding step that runs behind a chatbot whenever it speaks about today. Brave has said its products handle more than 22 million AI answers a day.

The API is not the only Brave product the complaints go after. Ask Brave, which composes answers inside Brave Search off the same index, and the browser itself are both on the list. What the plaintiffs offer as evidence is the length of those answers. One Times article came back through the LLM Context endpoint with 1,291 of its 1,295 words intact, the complaint says. On the Ask Brave side, the plaintiffs point to a Chicago Tribune article reproduced at more than 1,200 words and a Daily News article at close to 600. That is nearer to the article itself than to a summary of a search result, the plaintiffs argue.

The complaints come at one business from five angles. The business runs as a sequence: crawl to collect, index to store, sell to third parties, and earn directly through the browser. Looking at where each claim attaches shows why this is not an ordinary unauthorized-copying case.

• Direct copyright infringement — the claim that scraping articles into the index is itself an act of copying. It targets the front of the chain.
• Contributory and vicarious infringement — the claim that selling the index through an API helped AI companies infringe and that Brave profited from doing so. These target the sales stage.
• DMCA circumvention — brought under sections 1201(a)(1)(A) and (a)(2) of the US Copyright Act. The plaintiffs say the paywall-blocking feature in Brave browser is primarily designed to defeat the access controls on paid articles, and under section 1203 they ask for the destruction of the devices and means involved.
• Unjust enrichment — a fifth claim brought under New York state law. The complaint states that it stands whether or not copying or distribution is established.

The unjust enrichment claim stands apart because of Brave's advertising business. Brave browser blocks the ads on a web page by default. Into the space where the publisher's ad used to run, Brave places its own, and returns 70 percent of the revenue to users as Brave Rewards. Brave and the advertiser keep the rest. As a yardstick, the complaint cites the $114 million in digital advertising revenue that The New York Times reported for the second quarter of 2026.

The same structure explains why paywall circumvention and AI data sales appear in one complaint. A feature that climbs the wall around paid articles improves the index, and a better index sells for more. To the plaintiffs the two are not separate acts but the front and the back of one business.

Anthropic, Perplexity, Mistral, xAI and Cohere appear in the complaints by name, as customers of the Brave Search API. None of the five is a defendant. The Times already has comparable suits running against OpenAI and Microsoft, filed in 2023, and against Perplexity. This time it stepped past the companies that build and serve the models and picked the stage before them.

The diagram below carries over the three boxes the complaints draw. On the left are the publishers who make the articles, in the middle is the Brave index that scraped them, and on the right are the AI companies that buy from it. If AI copyright suits have so far clustered at the right-hand end, the seat in the dock this time belongs to the middle box.

Perplexity is the only one of the five to speak. Jesse Dwyer, the company's chief communications officer, told The Verge: "The New York Times is overpaying its attorneys and underpaying its journalists. (Who could've easily identified the disinformation in this suit.)" The rebuttal stopped there. Brave and the other four did not respond to requests for comment, and as of November 7 no public statement from Brave and no answer filed with the court had appeared.

Silence does not leave Brave's legal position unknown. Its deadline to answer under the Federal Rules of Civil Procedure has not run out, and the company has already carried the same question into another courtroom on its own initiative. In March 2025 Brave sued News Corp parties in the Northern District of California, seeking a declaration that indexing and summarizing are fair use and the ordinary work of any search engine. That case was dropped in June 2025 and refiled in May 2026 after licensing talks broke down, as 4:26-cv-04758-JST. In July, News Corp countersued, calling the conduct brazen theft. Among the things the counterclaim asks for is the destruction of the index and the databases.

The two complaints in Manhattan also ask for more than damages and disgorgement of profits; they seek a permanent injunction. Customers being left off the defendant list does not leave them untouched. If the seller has to carve parts out of the index or drop particular domains, the first thing to wobble is the answer quality of every service grounded on it. Contract terms that change, or supply that stops, work the same way. Having your name written into a complaint can itself become a term at the next negotiating table.

When a team builds an AI service, a search API is close to a part you bolt on without much thought. You compare latency, unit price and coverage, pick one, and assume the supplier took care of where the data came from. That assumption is exactly what this case touches. The legality of an index stays an internal matter for the supplier right up to the moment it reaches a courtroom, and from there it becomes the problem of every service that bought from it.

Writing down data lineage has been treated as a cost paid for audits and reproducibility. In a case like this, lineage does something else. Nothing but that record backs up the sentence "we believed the supplier collected it lawfully." Which API you used, from when, on what terms, and which feature the output went into: those entries are what let a company speak about a narrow scope. Without them it stands on the side where everything is in doubt.

For anyone using a search or grounding API, there are four things worth getting on paper before the next renewal.

• How was the supplier's index collected? Ask whether the crawler identifies itself and whether publishers have a channel for saying no.
• Does content from behind paid subscriptions enter the index? If it does, is the basis for that written into the contract?
• Where does liability land if a copyright dispute arises? Does the indemnity reach back to how the supplier obtained the source data?
• Which features of your own service lean on that API? Do you have a list of what stops when the supply does?

None of the four asks about the accuracy of the data. They ask where it came from and whose hands it passed through. A lawsuit landing on the middle link of the supply chain is under way in two courts at once, counting Brave alone. Writing the lineage down costs less than the bill that arrives at that point.

Thanks for reading this far. The facts of the case are set out in the reporting by TNW and CyberInsider, and Brave's search business is covered in the Search Engine Journal piece. If you have ever written data provenance or indemnity into a contract with a search API supplier, we would like to hear how that sentence came out.

Lawsuit Coverage

• 1.Dillet, R. (2026). "New York Times and other outlets sue web browser Brave over alleged content scraping." TechCrunch.
• 2.PPC Land. (2026). "New York Times and three papers sue Brave over AI answers and ad blocking."
• 3.AI Weekly. (2026). "NYT, Tribune Papers Sue Brave Over 30B-Page AI Scraping Index."
• 4.The Next Web. (2026). "New York Times and three other US papers sue Brave over AI scraping."
• 5.Search Engine Journal. (2026). "NY Times Accuses Brave Of Scraping And Exploiting Content."
• 6.CyberInsider. (2026). "New York Times Brave Lawsuit: Paywall Bypass, AI Scraping."

Brave's Own Record & 2023 Independence Background

• 7.Brave Software. (2023). "Brave Search removes last remnant of Bing from search results page." Brave Blog.
• 8.Mehta, I. (2023). "Brave Search doesn't use Bing's index anymore." TechCrunch.
• 9.Brave Software. (2023). "Brave releases its Search API, bringing independence and competition to the search landscape." Brave Blog.
• 10.Brave Software. "Brave Search API." brave.com.