Cloudflare Accuses Perplexity of Using AI Bots to Crawl Blocked Websites in Secret

AI Startup Accused of Circumventing Website Restrictions
Cloudflare's recent investigation alleges that AI search company Perplexity has been employing tactics to bypass crawling restrictions implemented by website owners. The internet infrastructure provider reports observing systematic attempts to disguise Perplexity's web crawlers when encountering access barriers.
The Circumvention Tactics
According to Cloudflare's findings, Perplexity's crawlers initially present standard identification ("PerplexityBot" or "Perplexity-User"). However, when blocked through:
- robots.txt directives
- Web Application Firewall rules
- Other access restrictions
The system allegedly switches to masking its identity as a regular Chrome browser user on macOS, utilizing:
- Rotating IP addresses not officially listed
- Changing autonomous system network identifiers
- Undocumented user agent patterns
Scale of Activity
Cloudflare documents this behavior across:
- Tens of thousands of domains
- Millions of daily requests
- Various network configurations
Company Responses
Perplexity's official statement contests Cloudflare's characterization, describing it as:
- A "publicity stunt"
- Containing "many misunderstandings"
- Potentially confusing legitimate user traffic with scraping activity
The startup attributes some detected activity to:
- Actual users making specific requests
- Third-party service BrowserBase
- Occasional technical necessities
Industry Context
This incident follows:
- Previous reports of Perplexity bypassing paywalls
- The company's past attributions to third-party crawlers
- Growing industry concerns about AI content scraping
Cloudflare has taken action by:
- Removing Perplexity's verified bot status
- Implementing new blocking measures
- Expanding default AI crawler restrictions
The situation reflects broader tensions between:
- AI companies' data needs
- Publisher rights and protections
- Evolving internet infrastructure responses
Related article
Google Introduces 9 Exciting New Features in Home App Soft Launch
Google Home App Unveils Major Updates in 2025 PreviewGoogle has significantly enhanced its Home app with several groundbreaking features currently available in public preview. Smart home enthusiasts can now enjoy Nest Cam's picture-in-picture functio
Grammarly Expands into AI-Powered Productivity Platform
Grammarly has announced plans to acquire the popular email productivity app Superhuman, according to an official statement. The move strategically aligns with Grammarly's existing email optimization features, which currently assist professionals in r
Jony Ive’s Secretive OpenAI Device Reportedly Ditching Screens
The enigmatic collaboration between OpenAI and renowned designer Jony Ive is developing a pocket-sized, context-aware device that notably won't include screens or take the form of eyewear. Internal communications obtained by The Wall Street Journal r
Comments (0)
0/200

AI Startup Accused of Circumventing Website Restrictions
Cloudflare's recent investigation alleges that AI search company Perplexity has been employing tactics to bypass crawling restrictions implemented by website owners. The internet infrastructure provider reports observing systematic attempts to disguise Perplexity's web crawlers when encountering access barriers.
The Circumvention Tactics
According to Cloudflare's findings, Perplexity's crawlers initially present standard identification ("PerplexityBot" or "Perplexity-User"). However, when blocked through:
- robots.txt directives
- Web Application Firewall rules
- Other access restrictions
The system allegedly switches to masking its identity as a regular Chrome browser user on macOS, utilizing:
- Rotating IP addresses not officially listed
- Changing autonomous system network identifiers
- Undocumented user agent patterns
Scale of Activity
Cloudflare documents this behavior across:
- Tens of thousands of domains
- Millions of daily requests
- Various network configurations
Company Responses
Perplexity's official statement contests Cloudflare's characterization, describing it as:
- A "publicity stunt"
- Containing "many misunderstandings"
- Potentially confusing legitimate user traffic with scraping activity
The startup attributes some detected activity to:
- Actual users making specific requests
- Third-party service BrowserBase
- Occasional technical necessities
Industry Context
This incident follows:
- Previous reports of Perplexity bypassing paywalls
- The company's past attributions to third-party crawlers
- Growing industry concerns about AI content scraping
Cloudflare has taken action by:
- Removing Perplexity's verified bot status
- Implementing new blocking measures
- Expanding default AI crawler restrictions
The situation reflects broader tensions between:
- AI companies' data needs
- Publisher rights and protections
- Evolving internet infrastructure responses
Google Introduces 9 Exciting New Features in Home App Soft Launch
Google Home App Unveils Major Updates in 2025 PreviewGoogle has significantly enhanced its Home app with several groundbreaking features currently available in public preview. Smart home enthusiasts can now enjoy Nest Cam's picture-in-picture functio
Grammarly Expands into AI-Powered Productivity Platform
Grammarly has announced plans to acquire the popular email productivity app Superhuman, according to an official statement. The move strategically aligns with Grammarly's existing email optimization features, which currently assist professionals in r
Jony Ive’s Secretive OpenAI Device Reportedly Ditching Screens
The enigmatic collaboration between OpenAI and renowned designer Jony Ive is developing a pocket-sized, context-aware device that notably won't include screens or take the form of eyewear. Internal communications obtained by The Wall Street Journal r




