Court: Google can't use the DMCA to block scraping public results
SerpApi ruling leans on hiQ v. LinkedIn: scraping public pages for datasets, RAG or agents carries less copyright risk — though ToS, CFAA and GDPR still bite.
Copy markdown
What the court actually held
Chief Judge Yvonne Gonzalez Rogers (N.D. Cal.) dismissed Google's DMCA §1201 anti-circumvention suit against search-API firm SerpApi. Claims over plain, non-copyrighted search results were tossed with prejudice; only claims tied to licensed Knowledge-Panel images survived, and Google has 21 days to amend those.
The rule of thumb: the lock has to be on the book
The court held that an anti-bot gate over a 'mixed bag' of mostly-uncopyrighted data isn't a copyright access control — §1201 protects works, not whole websites. Blocking scrapers from factual results, the judge reasoned, is not circumvention. The decision leans directly on hiQ v. LinkedIn.
Why it matters for your data pipeline
Scraping the public web underpins training sets, RAG corpora and browsing agents. This shrinks one of the scariest legal threats — a DMCA claim that carries statutory damages — for pulling public, factual data. Standing up a scraper or dataset this weekend just got a little safer.
The catch: this isn't open season
It's Ninth Circuit only, addresses just DMCA §1201, and leaves every other lever intact: Terms-of-Service breach-of-contract claims, the CFAA, and GDPR/privacy law all still apply. Wholesale copying of copyrighted content — images, articles — stays risky, and Google can still amend and appeal.