Published
Google Crawls the Web, Then Sues a Scraper
Google built Search by sending automated software across the public web, downloading pages, analyzing their content, and storing information in an index. On December 19, 2025, Google sued SerpApi over automated access to Google's public search results.
Our view is straightforward: Google is on the wrong side of this argument. A company whose core search product depends on automated access to other people's public pages should face serious scrutiny when it tries to reserve programmatic access to public search results for itself.
That does not mean every scraper behaves responsibly or that every form of automated access is lawful. It means the rules should be consistent, the facts should be stated accurately, and one dominant crawler should not get a moral exemption because it calls its scraping "indexing."
The policy symmetry is easier to see without the branding:
Public page -> automated fetch -> structured Google index
Public result -> automated fetch -> structured third-party data
That comparison does not erase legal differences around access controls or licensed content. It shows why the identity of the crawler cannot be the whole rule.
Google describes an automated, opt-out web crawler
Google's own guide to how Search works says Search is fully automated and that the vast majority of pages in its results are not submitted manually. Its crawlers discover pages, download text, images, and video, render JavaScript, analyze the material, and store information in Google's index.
The public owner does not receive an individual permission request before each fetch. Instead, Google documents controls that a site can use to object or limit access. Its robots.txt guide explains how to restrict crawling, while noindex, authentication, and removal tools address other visibility and access goals.
That is an opt-out framework. Google automatically visits publicly accessible pages unless a technical or policy signal tells it not to. The model helped create a useful search engine, and publishers often welcome the resulting traffic. It is still automated collection of content created and hosted by other people.
What Google alleged against SerpApi
Google's December 2025 announcement said it sued SerpApi for circumventing security measures around copyrighted content appearing in Search. The linked complaint alleged violations of the Digital Millennium Copyright Act tied to Google's SearchGuard system and licensed third-party content.
Those are Google's allegations, not established findings. SerpApi disputed them.
The procedural history matters. The federal court dismissed Google's original complaint on July 20, 2026. The dismissal did not end the dispute. Google filed an amended complaint in August, and SerpApi reported on August 25 that it had moved to dismiss the amended claims. SerpApi also links the court's dismissal order from its July case update.
As of August 30, 2026, there is no final merits decision resolving the amended case. Anyone claiming that Google definitively won or that all scraping was declared legal is getting ahead of the record.
Google's best argument deserves a fair hearing
The strongest version of Google's position is narrower than "scraping is bad."
Google says its crawlers identify themselves and honor standard directives. It alleges that SerpApi evaded a technical access measure at very high volume. It also says some Search surfaces contain material licensed from third parties, which raises different questions from collecting bare titles and links on an ordinary results page.
Those distinctions can matter legally and ethically. Crawling, scraping, indexing, copying, and circumventing an access control are overlapping concepts, not automatic synonyms. Authentication, contract terms, robots directives, copyright ownership, rate, purpose, and jurisdiction can change the analysis. Our web-scraping legal guide covers those considerations without pretending there is one universal rule.
The double standard remains
Google asks publishers to accept a default in which public pages may be found, downloaded, rendered, analyzed, and indexed automatically. Publishers who object must learn and deploy Google's supported controls.
When developers automate access to Google's public results, Google describes that access in entirely different moral terms and asks a court to stop it. Its lawsuit may turn on specific allegations about technical controls and licensed works, but the economic position is plain: automated collection built Google's search index, while outside automation of the resulting search pages threatens Google's control over access.
Calling the first activity "crawling" and the second "scraping" does not settle the policy question. The same basic act sits underneath both: software requests a public page and turns the response into structured information.
A defensible open-web policy should focus on conduct. Respect meaningful access controls. Do not obtain private data by deception. Limit load. Identify and honor rights in copied material. Give public facts, links, and interoperable data room to move. Apply those principles whether the crawler belongs to a startup or one of the largest companies on the internet.
A friendlier search API starts with the bill
Developers should be able to buy search data without a renewal cycle that dictates when credits reset or pressures them into a larger monthly commitment.
PrismCrawl pricing uses one-time prepaid packages. The smallest purchase is $5 at $0.30 per 1,000 successful searches. Larger purchases lower the rate to $0.15 per 1,000. Credits remain valid for 90 days, and there is no subscription or automatic renewal.
The difference is visible in the listed entry rates below. We checked each linked public pricing page on August 30, 2026.
| Service | Listed entry purchase | Listed price per 1,000 searches |
|---|---|---|
| PrismCrawl | $5 one-time | $0.30 per 1,000 successful searches |
| Bright Data | Pay as you go | $1.50 per 1,000 requests |
| SearchApi | $40 per month for 10,000 | $4.00 per 1,000 searches |
| SerpApi | $25 per month for 1,000 | $25.00 per 1,000 searches |
PrismCrawl's $0.30 entry rate is 80 percent below the lowest external rate in that table, Bright Data's $1.50 pay-as-you-go rate. The other listed entry gaps are larger. At volume, PrismCrawl's published rate falls to $0.15 per 1,000 successful searches.
This is a price comparison, not a claim that the services have identical engines, support, compliance programs, throughput, or feature breadth. Promotions and enterprise agreements can change the calculation. Check the linked pages for current terms before choosing a provider.
One successful request means one credit
PrismCrawl's billing unit is deliberately easy to audit:
- one successful search request consumes one credit across the documented search endpoints;
- a failed search consumes no credit;
- the complete API and SERP features are included at every pricing level;
- the response is live and uncached;
- there is no automatic monthly reset or renewal.
Success-only billing aligns our incentive with the customer's. Failed attempts do not consume a customer credit, so revenue depends on delivering successful searches. The PrismCrawl product page currently reports a 99 percent success rate. That is a service measurement, not a guarantee that every query will succeed.
The uncached part matters too. A search API can make a fast response look current by serving an earlier result. PrismCrawl runs the request live. Request-history retention stores completed artifacts so customers can inspect what happened; it does not turn a prior response into the answer to a new search.
The result is a simple cost model. A successful Google, Bing, Amazon, or Maps search is one credit. Errors are free. A news feature or another supported result type does not introduce a hidden multiplier. Buy a package when the work exists, use it within 90 days, and buy another only when needed.
Open access needs clear rules, not selective outrage
Google is entitled to protect private systems, enforce legitimate rights, and challenge conduct it believes breaks the law. It should also acknowledge the foundation beneath its own product: an automated crawler that continuously collects and organizes public web content at a scale almost nobody else can match.
Our objection is to the asymmetry. Google treats automated public-web access as socially valuable when Google does it and inherently suspect when an outside developer does it to Google. The pending SerpApi case will address narrower legal claims. The broader policy question belongs to everyone who publishes, searches, researches, and builds on the open web.
PrismCrawl's answer is practical. Provide live search data through a documented API, price it low enough for ordinary developers, charge only when a search succeeds, and avoid a renewal-centered plan. Try 25 searches free or inspect the exact request and response contract in the API reference.
Frequently asked questions
Does Google scrape the web without permission?
Google says its automated crawlers discover and download public pages, while site owners can use robots.txt, noindex, access controls, and removal tools. That is generally an opt-out model rather than a case-by-case permission request, but whether a particular automated access is lawful depends on the facts and jurisdiction.
Did Google win its lawsuit against SerpApi?
No final result had been reached as of August 30, 2026. The court dismissed Google's original complaint on July 20, Google filed an amended complaint, and SerpApi moved to dismiss the amended claims on August 25.
How does PrismCrawl charge for search requests?
Each successful search uses one credit across the documented search endpoints. Failed searches use no credit. Paid credits are purchased once, remain valid for 90 days, and do not renew automatically.
Does PrismCrawl return cached search results?
No. PrismCrawl runs live, uncached searches. Request-history retention lets customers inspect completed requests, but a retained artifact is not substituted for a future live search.