Brave Search is about to stop linking straight to websites. In a blog post on 11 November, Brave said that “in the upcoming days” every result link will go through a cloaking scheme: a search.brave.com/go/v1#… address with the real destination encrypted after the hash. It blames Google.
Google closed the window, so the scrapers came to Brave
Google started routing result clicks through google.com/goto earlier this year. Techaeris notes that Google confirmed this in late August, saying only that it has “a long history of deploying technical measures against evolving forms of abuse”. Brave’s account is blunt: “Google’s implementation did not so much abolish the scraping trade as relocate it.” Scrapers that resold Google results to chatbots and AI agents went looking for “a search engine that maintains an independent index rather than reselling someone else’s”. That means Brave.
Brave says the result is traffic spikes of “several thousand requests per second”. That forces tougher bot filtering, and real users pay for it with slower results and more CAPTCHAs. Brave also makes an interesting market argument. Google tolerated scrapers for years because cheap Google results “demonetized” rivals like Bing and Brave. Then chatbots arrived, and the same scrapers were effectively subsidising OpenAI, Grok and Meta. So Google closed the window.
How Brave’s version differs from Google’s
The key difference is privacy. Each link has its own key. When you click, your browser requests that key in disguised form, using a partially oblivious pseudorandom function (RFC 9497). It then decrypts the destination locally and navigates straight to the site. Brave says it can tell that someone opened a link from a given results page and when, but not which one. A person opens one or two results per page. A scraper wants all ten and has to request every key, which Brave can count per page. That is the defence.
Brave also lists the side effects. Copied links are redirects, not destinations, and may expire. Redirects can be abused for phishing, so Brave has added revocation, a time-to-live and an interstitial warning. Users with JavaScript turned off fall back to a conventional redirect, where Brave does see the clicked URL.
What this means for SEO
Three things.
- Rank tracking for Brave gets harder and more expensive. Any tool that reads Brave positions from the HTML will now see cloaked URLs. Mapping a position to a URL means fetching a key per result, and Brave is specifically counting those requests. Expect vendors either to pay for the Brave Search API or to quietly drop Brave tracking. If your Brave data goes strange next week, check the tool before you blame an update.
- Brave visibility matters more for AI answers than for Brave’s own clicks. We reported that 89% of Claude’s cited pages sat in Brave’s top 20 in Peec’s tests. Brave is pushing everyone who wants that data through its paid front door, which strengthens its hand with AI companies.
- Analytics should be fine, but check it. The browser decrypts and goes straight to your site, so the visit doesn’t pass through Brave’s servers. Brave hasn’t said exactly what referrer header sites will receive. Compare your
search.brave.comreferral numbers before and after rollout.
My take
Brave’s timing is as interesting as the cryptography. The company is being sued by the New York Times and Tribune papers over the Search API and Ask Brave. It just reported 2.71bn October searches. And it is now arguing in public that its independent index is the scarce asset the AI market wants. Cloaking links makes that asset harder to take for free, and nudges every scraper towards a paid API that Brave can meter, license and defend. It isn’t a wall, since determined scrapers will still pay the per-key cost. But it is a price signal, and one that Google’s goto move made possible.
As Brave puts it: “Anyone who wants our search results at scale is welcome to use the front door.”
