# Google's Sign-In Test Turns a CAPTCHA Wall Into an Identity Wall > Google is testing a sign-in wall on deep search results weeks after a DMCA loss to SerpApi, swapping an anonymous CAPTCHA for an identity check. Canonical: https://brandonlazovic.dev/articles/google-scraper-wall-after-serpapi/ Author: Brandon Lazovic Published: 2026-08-14 ## The short version - Google is testing a sign-in requirement on deep search results a few weeks after a court dismissed the core of its DMCA lawsuit against scraper SerpApi, and the pattern that matters is platforms turning to technical enforcement once a legal claim fails, not this one test. - The sign-in test rests on a single screenshot reported by Search Engine Land with no confirmation from Google and no evidence of a wider rollout, so treat it as an unconfirmed sighting. - Google already interrupts deep search paging with a CAPTCHA, so the real change would be a swap from an anonymous puzzle a scraper can solve at scale to an identity Google can suspend or trace across sessions. - The court permanently dismissed Google's DMCA claim over uncopyrighted search-result data, but gave Google until August 10, 2026 to refile a narrower claim tied to copyrighted Knowledge Panel content, so the case is not fully closed. Google appears to be testing a sign-in requirement for searchers who page deep into search results, based on one screenshot a searcher posted and Search Engine Land reported on August 4, 2026. [1] Because Google has not confirmed the test, described its scope, or said whether it extends past this single sighting, the right label for it is observed rather than confirmed. Two weeks earlier, a federal court had dismissed the core of Google's DMCA lawsuit against SerpApi, a company built on scraping and reselling Google's search results. [2] Google has not linked the two events, and neither does this piece. Look past both sightings and one pattern remains underneath. Whether or not this exact test ever ships broadly, a platform that loses a legal fight over unauthorized access tends to reach for a technical control instead, and anyone whose SEO tooling depends on reading results past page one should track that shift. > August 10, 2026: the deadline for Google to refile the narrower slice of DMCA claims against SerpApi the court dismissed with leave to amend, itself a sign the case is not over. ## What did Google actually test, and how confirmed is it? Search Engine Land reports that Google is running what it calls "a limited test" asking a searcher to sign into a Google account, rather than solve a CAPTCHA, after paging past the first few pages of results. [1] In the one screenshot behind the story, posted on X by a searcher named Kamlesh Shukla, the page reads "Sign in to verify you're a human and see more results." [1] The chain of evidence is thin. Search Engine Land's own reporting says plainly that Google has not commented and that it is unclear whether the test will roll out more widely. [1] One person, one screenshot, one outlet, zero confirmation from Google: nothing that follows should be read as more certain than that chain allows. ## Why did Google's DMCA case against SerpApi collapse, and is the ruling actually final? On July 20, 2026, a federal judge dismissed the core of Google's lawsuit against SerpApi, and one part of that dismissal is permanent, with no chance to refile. Google had sued under the DMCA, the Digital Millennium Copyright Act, leaning specifically on its anti-circumvention rule, the provision that makes it unlawful to bypass a technical barrier guarding copyrighted material. [2] [3] Because that rule does not protect access to material the statute never covered in the first place, Judge Yvonne Gonzalez Rogers, of the US District Court for the Northern District of California, granted dismissal "without leave to amend" for every claim resting on search results that carry no copyrighted content. [3] That is the part of the ruling that matters for ordinary rank-checking and SERP-scraping, because plain organic results, URLs, and snippets are exactly the uncopyrighted material SerpApi's business runs on. The same order is not a full loss for Google, and treating it as one misses its actual scope. For the narrower slice of claims resting on search results that do carry a copyrighted component, chiefly Knowledge Panels built from licensed third-party content, the court granted the motion "with leave to amend," giving Google 21 days from the order to refile if it can show the copyright holders authorized Google's protection measure. [3] Twenty-one days from July 20 lands on August 10, 2026. The public docket showed no amended complaint filed as of its last recorded update, July 29, 2026, so this narrower claim is pending rather than resolved, and a reader checking this after publication should look at the live docket rather than take that date as settled. [4] The order also settles a question this piece needs answered before leaning on any CAPTCHA comparison. Because the court accepted the complaint's account for the purpose of the motion, it found enough there to infer that SerpApi masked automated queries and solved SearchGuard's JavaScript challenge, sufficient to establish that circumvention occurred. [3] The claim still failed, for a narrower reason. Only when a barrier guards a copyrighted work with that owner's authority does circumventing it break the DMCA, and a page of factual search snippets is not one. [3] A barrier the law will not protect, sitting next to a barrier that still works technically, is exactly the space an identity check would fill without needing the DMCA, or any court, at all. That connection is my own inference; Google has not confirmed any such motive. SerpApi's chief executive, Julien Khaleghy, called the outcome "a win not just for SerpApi, but for all who depend on an open internet." [5] Treat that framing as the defendant's own victory lap; SerpApi just won the larger half of the case it was fighting. ## What kind of wall did Google already have before this test, and what actually changes? Google already interrupts a search session that pages too deep with a CAPTCHA, and that fact gets lost in coverage that frames the sign-in test as Google blocking scrapers for the first time. [1] Normally, per Search Engine Land's own report on the sighting, the searcher who hit the sign-in wall would have been served a CAPTCHA at that point in the session, not a request to sign in. [1] SerpApi's own engineering blog documents the same wall from the vendor's side. Per the company's own words, its API calls carrying the num parameter, the setting controlling how many results come back per request, have "historically been more prone to CAPTCHA challenges." [7] None of this is a new defense. Solving CAPTCHAs to page deep into Google is what a scraping operation has always had to do to stay in business. Narrower than "blocking scrapers" is what the sign-in test would actually change, if it ships beyond this one sighting. It swaps what kind of proof the wall demands. A CAPTCHA verifies that something solved a puzzle. It says nothing about who or what solved it, and a solve can be bought, automated, or resold at scale for a fraction of a cent. A signed-in Google account instead verifies an identity Google already holds data on, one it can suspend or trace back to a person or organization across every future session. Connecting the ruling to the test this way is my own read, not Google's. The company has not confirmed anything about its own motives for the sign-in wall. None of this required a lawsuit to be against the rules. Under Google's own spam policy for web search, scraping results for rank-checking, or any other automated access without express permission, already counts as "machine-generated traffic," and Google says plainly that "such activities violate our spam policies and the Google Terms of Service," with demotion or removal as the consequence, independent of any CAPTCHA or sign-in wall. [6] The DMCA case was always the narrower, harder claim, turning on whether evading SearchGuard counts as illegal circumvention of a copyright-protection measure rather than on whether scraping breaks Google's rules at all. Google already had the ordinary lever. It reached for a copyright-law lever anyway, lost most of that fight, and appears to be testing a third lever now, one that needs neither a court nor a takedown notice to work. ## Why would an identity check change the economics of scraping differently than a CAPTCHA does? An identity check changes the economics because Google can act on an account across every future session, while a CAPTCHA solve is discarded the instant it succeeds. A CAPTCHA-solving service sells anonymous, disposable proof: pay a few cents, get a token, and the token carries no history back to whoever purchased it. A Google account carries a persistent identity Google already controls. It can require phone verification and shut an account down the moment its behavior pattern looks automated, and that lever keeps working across sessions in a way a CAPTCHA token never does. Think of a CAPTCHA solve as a fake ID checked once at the door, and a signed-in account as a membership card checked against a name on file. The first stops a forgery in the moment; the second can bar that name the next time it shows up. The analogy breaks at scale, because minting a new membership is far cheaper online than forging a new physical identity, so the wall raises cost and leaves a trail rather than closing the door outright. At scale, buying or renting real, aged Google accounts is not impossible; account farms already exist for other purposes, such as fake reviews. It is a different cost structure than buying disposable CAPTCHA solves, and it leaves an identity trail a CAPTCHA solve does not, so the likely effect is to raise the price and the exposure of scraping past page one rather than to make it categorically impossible. That is a reasoned claim about incentives, not a tested result. ## How much of the SEO measurement industry actually depends on reading results past page one? Two categories of SEO measurement structurally depend on reading Google's results well past the first page, and a third does not. Share-of-voice modeling and competitive visibility tracking both need many tracked positions across a category to compare footprint over time, so both break the moment a wall stops a crawl before it reaches those positions. On one head term, a single rank check carries a much lighter dependency, usually needing only the first page or two, the range most dashboards already surface without a fight. SerpApi's own team names exactly who feels a restriction on deep access first. Its engineering blog says SEO practitioners "oftentimes rely on getting as many results as possible in as few searches as they can," and that a separate Google change limiting results per request "makes it harder to analyze results quickly at scale and requires them to paginate." [7] That describes a different Google change, not the sign-in test, but it is the clearest public admission from inside the measurement industry that its own tooling assumes deep results stay available on demand. No client dataset backs the structural split above, and no percentage in this piece should be read as measured. From building and auditing this tooling, the split is a practitioner's read on where the dependency sits, reasoned rather than counted. ## What should you check on your own tooling right now? Open a fully signed-out browser window, run one of the queries your rank tracker or SOV tool already monitors, and click forward past page three. If a CAPTCHA already stops you there, your tool clears that same wall every day, and a sign-in requirement would upgrade it from a puzzle a script can solve to a check a script cannot pass without a real, traceable account behind it. Ask your vendor directly which positions they actually claim to cover, one page deep or several, and what happens to a report when a crawl gets challenged mid-session. With a specific fallback mechanism, a vendor is already managing this risk. One with only reassurance is the vendor whose numbers move first if this test ever ships broadly. None of this needs the sign-in test to survive contact with a wider rollout, and it may well not; Search Engine Land itself calls it a limited test with no confirmation of anything past the one sighting. [1] The larger pattern does not depend on that survival. Over seven months, Google lost the broad half of a legal argument that scraping factual search results violates copyright law, kept a narrower copyrighted-content claim alive past that loss with a refile deadline still open as of this writing, and now appears to be probing a technical control that needs neither the DMCA nor a court. [2] [3] [4] I covered a related version of the same challenge-screen mechanism from the site owner's side of this exact problem in [an earlier piece on bot-challenge screens and deindexing](/articles/google-challenge-screen-deindexing/). This time the mechanism points at the industry that measures search itself, and any tool assuming deep results stay freely readable should price in that risk now, while the test is still just a test. ## Sources 1. Search Engine Land: Google Search testing forcing searchers to sign in to get more search results (Barry Schwartz, Aug. 4, 2026) — https://searchengineland.com/google-search-testing-forcing-searchers-to-sign-in-to-get-more-search-results-484223 2. Search Engine Land: Google loses key DMCA claims against SerpApi in scraping lawsuit (Danny Goodwin, Jul. 22, 2026) — https://searchengineland.com/google-loses-key-dmca-claims-against-serpapi-in-scraping-lawsuit-483185 3. US District Court, N.D. Cal.: Order Granting Motion to Dismiss, Google LLC v. SerpApi, LLC, No. 4:25-cv-10826-YGR (Jul. 20, 2026) — https://storage.courtlistener.com/recap/gov.uscourts.cand.461513/gov.uscourts.cand.461513.42.0.pdf 4. CourtListener: Docket, Google LLC v. SerpApi, LLC, 4:25-cv-10826 (N.D. Cal.) — https://www.courtlistener.com/docket/72059948/google-llc-v-serpapi-llc/ 5. SerpApi: Google v. SerpApi, The Court Granted Our Motion to Dismiss (Julien Khaleghy, Jul. 21, 2026) — https://serpapi.com/blog/google-v-serpapi-the-court-granted-our-motion-to-dismiss/ 6. Google Search Central: Spam policies for Google web search, machine-generated traffic — https://developers.google.com/search/docs/essentials/spam-policies#machine-generated-traffic 7. SerpApi: Google Experiments With Restricting Results Per Page (Bartek Marmolowski, Sep. 16, 2025) — https://serpapi.com/blog/google-experiments-with-restricting-results-per-page/