SerpApi's workaround blocked: Google restricts Light Fast API to three results

SerpApi encountered a significant technical obstacle three days ago when Google restricted the company's Google Light Fast API to return only three organic search results. This development marks the second disruption affecting the search API provider since September 14, 2025, when Google eliminated the num=100 parameter that previously allowed retrieval of 100 search results in a single request.

The GitHub issue tracking the problem shows multiple user reports documenting the limitation. One user named fadupla commented two days ago, stating the impact affected 282,000 API calls to Google Search Engine Results API in a single day. Multiple customer service representatives from SerpApi acknowledged additional reports through their support system, indicating widespread disruption across their client base.

SerpApi had launched the Google Light Fast API as a solution to the September parameter elimination. The endpoint provided access to 100 organic search results through a single request by focusing exclusively on organic results without rich features. According to the documentation, the API was designed to offer faster response times compared to SerpApi's regular Google Search API by excluding elements like knowledge graphs and AI Overview answers.

The timing of Google's restriction appears deliberate. The blocked endpoint had been operational for approximately two weeks before Google implemented the limitation. User ggennrich speculated 17 hours ago on GitHub that SerpApi was "eating the cost with this API and doing 10 queries to get 100 results for organic only and eliminate the computer power required to parse the rest of the rich results." The comment suggests Google identified the workaround as using "a separate endpoint that Google has since blocked."

The restriction to three results creates significant operational challenges for SEO platforms and marketing professionals who depend on comprehensive search result data. Google eliminated the num=100 parameter on September 14, 2025, fundamentally transforming how tools access search result data by forcing 10 separate requests instead of one to retrieve 100 results.

The parameter had been essential for comprehensive keyword research, competitor analysis, and large-scale search engine optimization strategies. Marketing professionals require detailed analytics to distinguish between campaign performance issues and algorithmic impacts on organic visibility, particularly during periods of high search result volatility in 2025.

Google's June core update caused dramatic ranking changes across 16 days, while April volatility reached high levels with particularly intense fluctuations between April 21-23. These algorithmic changes increased demand for comprehensive search monitoring tools precisely when access costs increased substantially through the parameter elimination.

SerpApi's documentation described the Light Fast API functionality before the restriction. The service accepted standard parameters including search query, geographic location, localization settings, and pagination controls. The num parameter could be set to values including 10, 40, or 100 to define the maximum number of results returned. The documentation warned that using num may introduce latency or prevent inclusion of specialized result types.

The API returned structured JSON data containing organic results with position, title, link, displayed link, snippet, and favicon for each result. Search metadata included status indicators flowing through Processing, Success, or Error states. The system provided cached results free of charge when queries and parameters matched exactly, with cache expiration after one hour.

Technical constraints affect how SEO platforms can respond to these restrictions. Rate limiting prevents rapid sequential requests that might replicate bulk data collection capabilities. Google's infrastructure monitors request patterns to identify automated scraping attempts. Distributed request strategies across multiple IP addresses or geographic regions may provide partial solutions but introduce additional complexity and infrastructure costs.

API-based alternatives through Google's official channels exist but typically provide limited result sets compared to direct SERP scraping. The Google Search Console API offers performance data but lacks comprehensive competitor visibility that many SEO tools require. Custom search solutions through Google's programmable search engine provide structured access but impose different limitations and cost structures.

The marketing community faces compounding challenges from multiple data access restrictions. ChatGPT referral traffic dropped 52% starting July 21, 2025, as citation patterns shifted dramatically toward Reddit and Wikipedia. SEO consultant Lily Ray theorized on October 1, 2025, that LLM tracking tools reported ChatGPT stopped citing Reddit, Wikipedia, and other sites as frequently beginning around September 11.

Ray's analysis posted at 12:04 PM on October 1 drew explicit connections between Google's parameter removal and ChatGPT's citation behavior. She noted the date coincides with Google removing the num=100 parameter, stating: "This would mean ChatGPT could no longer cite URLs ranking in 20-100, including Reddit and Wikipedia (two sites it tends to cite most frequently)." Ray concluded her post by asking whether the theory seems plausible, inviting industry feedback on the correlation.

The speculation gained traction within the search marketing community. Brodie Clark responded 55 seconds after Ray's post, noting that SerpApi was supposedly a leaked provider of ChatGPT search functionality. Clark referenced the workaround solution, pointing to SerpApi's blog post about retrieving top 100 results, and suggesting monitoring LLM tracking tools for citation pattern changes in coming weeks.

Ross Stevens added perspective 58 minutes after Ray's original post, questioning whether the citation decrease could happen so dramatically given that ChatGPT's core training data tends to be several months old. Stevens suggested the impact would potentially affect only the fresh data portion retrieved in real-time. John Morabito expressed full agreement 36 minutes after the post, stating the parameter change explained why citations dropped, though noting the Wikipedia and Reddit specifics might need additional investigation.

The timing alignment between Google's September 14 parameter elimination and the September 11 citation pattern shift creates an intriguing correlation. If ChatGPT relied on search APIs using the num=100 parameter to retrieve comprehensive result sets, the parameter's removal would directly limit which URLs the system could cite. Pages ranking beyond position 10 in traditional search results would become invisible to ChatGPT's citation selection process without multiple sequential requests.

The citation consolidation benefits specific platforms significantly. Reddit citations increased 87% starting July 23, reaching more than 10% of all ChatGPT citations. Wikipedia's citation share rose 62% to nearly 13% of total citations. The top three domains—Wikipedia, Reddit, and TechRadar—now control 22% of all citations, representing one in five ChatGPT citations going to just three sites.

Professional commentary on the situation emphasizes technical challenges ahead. Brodie Clark noted on social media that SerpApi was supposedly a leaked provider of ChatGPT search functionality. Clark referenced SerpApi's workaround solution in his October 1 comment, stating the company "found a workaround" to get top 100 results, adding that monitoring LLM tracking tools would reveal whether citation patterns change in coming weeks.

John Morabito attributed the citation decrease entirely to the Google parameter change, stating with certainty that the modification led ChatGPT to stop providing as many citations because the system cannot access as much of the search result set. The analyst argued ChatGPT prefers not to give citations when comprehensive result data becomes unavailable.

The industry-wide disruption affects SEO platforms differently based on their business models and technical capabilities. Semrush confirmed significant operational impact on September 14, acknowledging the modification forces tools to execute 10 separate requests instead of one to retrieve 100 search results, multiplying operational costs by a factor of 10.

The company publicly stated: "This is an industry-wide issue that impacts all rank tracking tools, but the good news is that we've already rolled out workarounds to minimize disruption and keep core functionality intact." Semrush emphasized transparency, noting users rely on ranking data for business operations. The platform implemented interim solutions while developing longer-term approaches to the cost increase challenge.

SISTRIX modified its data collection methodology after the parameter elimination, implementing changes progressively between September 17 and September 22. The company announced discontinuation of desktop SERP data updates and desktop visibility index on September 22. SISTRIX prioritized high-value keywords for deep measurement while implementing selective depth tracking for other terms.

Enterprise platforms serving large clients may absorb increased operational expenses more effectively than smaller providers targeting individual users or small businesses. Subscription pricing models across the SEO industry may require restructuring to accommodate higher data acquisition costs. Marketing teams may reduce their SEO tool subscriptions or consolidate services to manage budget pressures.

The cost increase may accelerate market consolidation within the SEO tools sector. Larger platforms with diverse revenue streams can better absorb operational cost increases compared to specialized providers focused solely on search result analysis. Alternative data sources become more attractive as Google's direct access costs increase.

SerpApi indicated on its public roadmap that the company is "actively exploring solutions" to the three-result limitation. The status label shows "prioritized" and "bug" classifications, suggesting the company treats the restriction as a technical issue requiring urgent resolution. Multiple team members added comments tracking user reports through customer support channels.

The GitHub thread demonstrates the restriction's immediate impact on production systems. One user reported launching 282,000 calls that experienced problems. Another customer service representative noted additional user interest in the num=100 parameter issue through separate support tickets. The volume of reports indicates SerpApi's client base relies heavily on bulk result retrieval for business operations.

Technical analysis from community members suggests limited pathways forward. The comment warning about "a game of cat and mouse" reflects industry awareness that workaround solutions may face continued restrictions. Google's ability to identify and block alternative endpoints demonstrates active monitoring of data access patterns that deviate from standard user behavior.

The restriction's implications extend beyond immediate technical challenges. AI search developments continue reshaping how users consume search information. With 58% of users encountering AI-generated summaries in March 2025, traditional search result monitoring faces additional relevance challenges beyond cost considerations.

The convergence of multiple access restrictions creates compound effects for marketing professionals. Reduced access to comprehensive search results coincides with changing user behavior patterns favoring AI-generated summaries. Citation pattern shifts in ChatGPT demonstrate how AI systems increasingly mediate information discovery, potentially reducing direct website traffic.

Professional tools attempting to maintain service quality face technical constraints preventing rapid sequential requests. Google's infrastructure actively identifies automated scraping attempts through pattern analysis. Geographic distribution strategies may provide partial solutions but require additional infrastructure investment and operational complexity.

The absence of official Google channels providing equivalent data access forces continued reliance on scraping methodologies. Custom search solutions through Google's programmable search engine impose different limitations that may not support specific data points required for competitive intelligence. The Google Search Console API lacks comprehensive competitor visibility essential for many analysis use cases.

Marketing technology vendors developing integration capabilities must now account for unstable data access patterns. Predictable data retrieval formed the foundation of automation workflows and reporting systems. Disruptions to access methodologies require architectural changes beyond simple cost adjustments.

The situation demonstrates platform power dynamics in digital marketing infrastructure. Google's unilateral decisions about data access affect entire industry segments dependent on search result information. The absence of advance notice for the September 14 parameter elimination left tools scrambling to implement workarounds under production pressure.

Alternative search engines may benefit from Google's data access restrictions. Marketing professionals seeking comprehensive search monitoring across multiple platforms may diversify their competitive intelligence approaches. Social media monitoring and content discovery platforms could gain market share as businesses reduce reliance on Google-specific tools.

The three-result limitation effectively eliminates the Light Fast API's utility for professional SEO applications. Comprehensive keyword research requires visibility across at least the first 100 results to identify ranking opportunities and competitive positioning. Three results provide insufficient data for strategic decision-making or performance tracking.

SerpApi's regular Google Search API remains available but requires multiple requests to retrieve 100 results, returning to the cost structure created by the original parameter elimination. The workaround's failure demonstrates Google's commitment to enforcing the new access limitations despite industry disruption.

Professional services dependent on search result data face strategic decisions about resource allocation. Increased operational costs may necessitate service restructuring or pricing adjustments. Some capabilities may become economically unfeasible under the new access constraints, forcing product portfolio changes.

The technical arms race between data access providers and platform restrictions shows no signs of resolution. Each workaround faces potential identification and blocking as platforms protect their infrastructure from automated data collection. The cat-and-mouse dynamic creates ongoing uncertainty for businesses building products on third-party data access.

Timeline

• September 11, 2025: LLM tracking tools report ChatGPT begins citing Reddit, Wikipedia less frequently, according to analysis by Lily Ray on October 1
• September 13-14, 2025: Google discontinues num=100 parameter without notice, limiting queries to 10 results instead of 100
• September 14, 2025: Keyword Insights announces parameter elimination at 7:45 AM; Semrush confirms industry-wide disruption; SISTRIX experiences initial delays with Visibility Index
• September 17, 2025: SISTRIX restores daily visibility data for DE, AT, CH, IT, ES, FR, UK, US, SE, BR
• September 19, 2025: SISTRIX restores data for NL, PL; fixes weekly rank tracking; DataForSEO implements new pricing structure
• September 22, 2025: SISTRIX discontinues desktop SERP data updates and desktop visibility index
• Mid-late September 2025: SerpApi launches Google Light Fast API as workaround solution
• September 30, 2025: SISTRIX CEO Johannes Beus publishes detailed update on response strategy
• September 28, 2025: Google blocks SerpApi's Light Fast API endpoint, limiting results to three
• September 29, 2025: User fadupla reports 282,000 affected API calls; SerpApi team acknowledges issue on GitHub
• October 1, 2025: Lily Ray posts theory connecting parameter removal to ChatGPT citation changes

Summary

Who: SerpApi, a search API provider serving SEO platforms and marketing professionals, experienced the restriction. Google implemented the blocking affecting SerpApi's client base including companies processing hundreds of thousands of daily search queries. Multiple enterprise SEO platforms including Semrush, SISTRIX, and Keyword Insights face operational challenges from the broader parameter elimination.

What: Google restricted SerpApi's Google Light Fast API to return only three organic search results instead of 100. The restriction blocked a workaround solution that provided bulk result retrieval after Google eliminated the num=100 parameter. The API had been operational for approximately two weeks before the blocking occurred. Multiple users reported the limitation through GitHub issues and customer support channels.

When: The restriction took effect three days ago, approximately September 28, 2025. Users discovered the limitation and reported it on September 29, 2025. This occurred two weeks after Google's September 14, 2025 elimination of the num=100 parameter that originally triggered the need for workarounds.

Where: The restriction affects the Google Light Fast API endpoint specifically. The issue manifests globally for all SerpApi users attempting to retrieve search results through the blocked endpoint. GitHub issue #3078 tracks the problem on SerpApi's public roadmap. The situation impacts SEO tools and marketing platforms worldwide that depend on comprehensive search result data.

Why: Google appears to have identified SerpApi's workaround as an alternative endpoint accessing search results beyond intended parameters. The restriction enforces Google's September 14 decision to eliminate bulk result retrieval. Marketing professionals require comprehensive search data for keyword research, competitor analysis, and performance tracking—capabilities that become economically challenging under the new access limitations requiring 10 separate requests for 100 results.