No, you cannot now scrape Google to infinity. That claim spread through SEO circles after the July headlines, alongside the opposite confusion, whether anybody can scrape Google at all anymore. Both misread the same order.
On 20 July 2026, Chief Judge Yvonne Gonzalez Rogers of the US District Court for the Northern District of California granted SerpApi's motion to dismiss Google's DMCA anti-circumvention claims. The DMCA is the US copyright statute, and its anti-circumvention section makes it unlawful to get around a technical lock that protects a copyrighted work. Google argued that SerpApi's scrapers circumvented SearchGuard, the anti-bot layer that blocks automated access to results. The court ruled, as Search Engine Journal reported, that blocking public search results is not DMCA circumvention when the results themselves contain no copyrighted content.
The ruling is DMCA-specific. It did not declare scraping Google legal in general, and the case continues in a narrower form.
What SearchGuard Is, from the Scraper's Side
The legal fight is over a technical layer, so it helps to know what the layer does. SearchGuard is Google's name for the defenses that decide whether a request for a results page came from a person or a program: rate limits on how many requests one address can make, checks on whether the requester looks like a real browser, and the CAPTCHA page that replaces the results when it decides the answer is no. To a scraper, hitting SearchGuard looks like a 429 error, a challenge page delivered with a 200 status, or a results page with the interesting parts missing. Getting around it, by rotating addresses, imitating browsers or solving challenges, is the circumvention Google said the DMCA forbids. The court's answer was that the DMCA only forbids circumventing a lock when there is a copyrighted work behind the lock, and a page of URLs, snippets and facts is not one.
What the Dismissal Did and Did Not Decide
| What it did | What it did not do |
|---|---|
| Dismissed the DMCA claims for results with no copyrighted content (URLs, snippets, facts) without leave to amend, which ends that theory for good | Did not give blanket permission to scrape everything, including licensed images or knowledge-panel content |
| Dismissed the DMCA claims for results carrying licensed images and knowledge-panel content because Google had not shown the copyright owners authorized it to deploy SearchGuard, with leave to amend, so Google can try again | Did not remove exposure under other theories: circumventing technical protections, breaching terms of service, or reusing genuinely copyrighted material can still create risk |
| Rejected SerpApi's argument that Google could not sue because it does not own its results, holding that DMCA protection is not limited to copyright owners | Did not end the case, and did not touch Reddit's separate DMCA suit against SerpApi over similar public-page scraping |
Two details show how narrow the order is. The court found Google had alleged enough facts to infer that circumvention of SearchGuard occurred, so the failure was legal rather than factual: SearchGuard cannot regulate access under the DMCA where there is no protected work behind it. And for the licensed content, the failure was authorization. Google had not shown it deployed SearchGuard with the copyright owners' permission, which the statute requires.
Why the Line Runs Between Public Facts and Licensed Content
The DMCA protects copyrighted works, not public data on its own. That is the whole boundary.
Plain search data, the URL, the snippet, the factual index entry, carries no copyrighted work for SearchGuard to guard, so bypassing SearchGuard to reach it is not DMCA circumvention under this order. Licensed material displayed inside results, such as licensed images and knowledge-panel text, is different. There a copyrighted work may exist, and the question becomes whether Google was authorized by the actual owners to guard it. On that question the court said Google had not made its showing yet, and gave it a path to try.
That is why "public facts safe from the DMCA, licensed extras still contested" is a more accurate summary than either headline reaction. One side asked whether the dismissal meant open season. The other declared there was no law left to stop scraping. The order says neither. The DMCA claim fails for uncopyrighted results and is paused for licensed ones.
What Is Still Open
Google was given 21 days to amend, with a deadline in early August 2026. The amended complaint targets licensed content, such as the Reddit data Google displays under license, and SerpApi has filed a new motion to dismiss that narrower complaint.
Separately, Reddit has its own parallel DMCA suit against SerpApi on similar public-page scraping questions. The 20 July order does not decide it.
Three questions are worth putting to any SERP vendor you pay while this is open. Does it fetch Google's pages itself, or license the data from someone who does, and does it say so in writing? What does it do with the licensed material inside results, the images and knowledge-panel text, which is exactly the content the amended complaint is about? And what does its contract say happens to you if its access is cut off or its method is ruled unlawful? A vendor with plain answers to all three is selling you a service. One that will not answer is selling you its exposure.
What It Changes If You Buy or Build
For anyone weighing a SERP API against scraping Google directly, the order moves one line and leaves the rest in place.
The DMCA is no longer the theory that reaches plain public results. That was the theory with teeth for a scraping vendor, and it is gone for URLs, snippets and facts. Everything else that made direct scraping expensive remains: the technical blocking itself, which Google enforces in layers regardless of what a court says about it, the terms of service you agree to, and copyright in whatever licensed material a result happens to carry. If your extraction touches images, knowledge panels or licensed text, the contested part of this case is your part.
The current shape, for anyone who arrived by searching whether scraping Google results is legal: the DMCA theory against plain public results was dismissed and cannot be brought again on that theory; the licensed-content theory was dismissed with a chance to refile and has been refiled in narrower form; the other theories were never decided by this order.
This article is general information about a US court order, not legal advice. Laws differ by jurisdiction and the case may change. For advice on your own situation, consult a lawyer.
