Google’s case against SerpApi is back, minus the claims the court barred it from re-pleading. On August 10, 2026, Google filed an amended complaint in Google LLC v. SerpApi LLC, Case No. 4:25-cv-10826-YGR, in the U.S. District Court for the Northern District of California’s Oakland Division, 21 days after Chief Judge Yvonne Gonzalez Rogers dismissed the original complaint in full. Gone for good are the claims covering search results that contain no copyrighted material: the court dismissed them without leave to amend, so Google cannot plead them again.
What the July Ruling Killed and What It Left Open
The July 20 dismissal split in two, by what the disputed search results contained. The permanently struck half rested on the court’s reasoning that a “compilation of publicly available links” lacks copyright protection. Claims over results built on licensed material were dismissed with leave to amend, for a narrower defect: Google had not adequately pleaded that SearchGuard, its JavaScript challenge against scraping tools deployed in January 2025, operated “with the authority of the copyright owner,” the standard set by 17 U.S.C. § 1201(a)(3)(B). SerpApi’s own read of the order, as Search Engine Land reported, was that the court “rejected Google’s attempt to expand the DMCA to control access to public pages.”
Four Categories of Authorization, One Licensor Named
Google’s amended complaint alleges four categories of authorization for SearchGuard. One licensor is named; two are not.
| Source of authority alleged | What it requires, per the complaint |
|---|---|
| An unnamed major partner, versions dating to 2017 | Requires “commercially reasonable efforts” to safeguard millions of licensed works from unauthorized third-party access |
| A second, unnamed provider | Licensed material “will not be available for download by third parties” |
| Reddit, the only licensor named in the complaint | Agreement directs Google “not to enable third parties to extract and independently commercialize the licensed content” |
| Google’s own Privacy Policy | Promises users that uploaded content, photos and reviews included, is protected from “unauthorized access” |
That characterization of the Reddit agreement is Google’s, per ppc.land’s coverage of the filing. Reddit is not a party to this case, and every word above describing its agreement comes from Google’s filing. Search Engine Journal’s summary of Google’s position is blunt: “To the extent it needs permission from copyright holders to run a measure like SearchGuard, the filing says, it had and has that permission.” Google’s complaint also alleges, per MediaPost, that licensing partners “expressly requested” the protection, and that Google receives what its own filing calls “hundreds of millions of artificial search requests each day” from SerpApi, an allegation in Google’s filing, not a measured count.
What Is Google’s Amended Complaint Against SerpApi About?
Google’s August 10, 2026 amended complaint in Google LLC v. SerpApi LLC (N.D. Cal., Case No. 4:25-cv-10826-YGR) re-pleads DMCA claims a judge dismissed for a pleading defect on July 20. It alleges SearchGuard, Google’s anti-scraping code, operates with copyright-owner authority because Reddit and two unnamed licensors instructed Google to protect their licensed content, and because Google’s Privacy Policy promises users their uploads are safeguarded. Nothing has been decided on the amended complaint, and SerpApi’s response is due within 14 days.
What Discovery Opens Up
The discovery stay lifts once any motion to dismiss is resolved, or SerpApi answers instead. That timing matters because Google built its amended pleading around the licensing agreements themselves, so contracts Google has kept confidential become discoverable. On ppc.land’s reading of the filing, once the stay lifts SerpApi can demand licensor identities, complete contract text, drafts, negotiation records, and the timing of SearchGuard’s design and deployment. Search Engine Journal notes that a second motion to dismiss faces procedural limitations, though none of the sources say how that plays out. SerpApi’s statement to MediaPost stayed short: “We remain confident in the strength of SerpApi’s position and look forward to responding.”
Whose Copyright Is Being Defended
Nothing here revives what the July order killed: on the court’s reasoning, a page of public search links with no copyrighted material in it cannot carry a DMCA §1201 claim. What Google was allowed to re-plead is a claim built entirely on someone else’s material, defensible, Google alleges, only because Reddit and two unnamed partners instructed it to protect that material, not because Google assembled the page it sits on.
That is a different instrument from two others covered here before. Publishers weighing whether to block Google Search outright are choosing among blunt opt-out buttons a publisher operates: snippet caps, a robots.txt line, a noindex tag. This ruling didn’t add to that list; it narrowed Google’s own DMCA remedy and hands publishers no new control of their own. Nor is it the mechanism behind the crawler-access dashboard Cloudflare built into beehiiv, which measures what every AI crawler takes and what it sends back and, on its paid tier, lets the publisher toggle them one by one. What Google alleges here is a private contract negotiated with a handful of partners, enforced through litigation, not a switch either side can throw.
The boundary this filing tries to draw around a search result is set, on Google’s own telling, by whoever holds the license upstream, not by the company that assembled the page. Making that argument put Google’s confidentiality in play: once the discovery stay lifts, contracts it kept private become material its opponent can demand in full.