Google Amends Lawsuit Against SerpApi Over Licensed Content

Aug 12, 2026 - 7:41 am 6 — by

Google Gavel

After Google's lawsuit against SerpAPI, the Google Search scraper tool, was dismissed a month ago, Google has decided to amend the complaint. Google filed the amendment on August 10, 2026 as an "Amended Complaint." Google is now taking an approach around protecting licensed content within its search results.

MediaPost wrote, "Google's amended complaint includes some new allegations regarding licensed content."

The example given was that Google alleges that Reddit, which licenses content to Google, specifically directed Google "not to enable third parties to extract and independently commercialize the licensed content." So here, Google is protecting its licensed content partners.

Google added in the amended complaint that its "licensing partners" have "expressly requested that Google prevent unauthorized access to licensed content."

Google is saying it has to safeguard its licenses and protect them from being scraped.

The official complaint should show up on Court Listener at the bottom soon? Update, here is the PDF of the amendment.

Forum discussion at X.

 

Popular Categories

The Pulse of the search community

Google Search Volatility

More Details

Search Video Recaps

 
Video Details More Videos Subscribe to Videos

Most Recent Articles

Search Forum Recap

Daily Search Forum Recap: October 21, 2026

Oct 21, 2026 - 10:00 am
Google Search Engine Optimization

Google Updates Control What You Share With Google Search With Existing Options

Oct 21, 2026 - 9:14 am
Google Updates

Google Search Ranking Volatility & Chatter Heats Up October 20 & 21

Oct 21, 2026 - 7:51 am
Google Ads

Google Ads New: Save & Test Asset Group Changes

Oct 21, 2026 - 7:41 am
Google Ads

Google Ads Updates Smart Bidding Learning Doc With 7-Day Limit & Best Practices

Oct 21, 2026 - 7:31 am
Google Ads

Google Merchant Center Specifies Fast Shipping Days By Country

Oct 21, 2026 - 7:21 am
Previous Story: Google AI Overview Local Results Pulling From Low-Quality Listicles