Unmatched Speed and Automated IP Recovery
This release significantly improves performance and stability across our scraping and batch data systems. We streamlined our request routing pipeline to shave off over two seconds of initial latency, and we hardened our batch poller and exporter tools to prevent duplicate executions or failed jobs. Additionally, we have rolled out interactive anti-bot difficulty scores and accessibility enhancements to our platform comparison pages.
The scraping engine now automatically identifies domains that only require background verification checks and skips unnecessary behavioral wait times. This cuts request latency on these sites by up to 4 seconds.
The platform now automatically detects connection-level proxy blocks and triggers a fresh egress route rotation. Instead of failing immediately when an IP is flagged, your requests are seamlessly retried with a clean IP.
We have added automatic detection and optimized routing pathways for multiple advanced, enterprise-grade bot protection systems. This ensures significantly higher success rates when scraping highly secured enterprise and retail websites.
New Features
5Faster passive verification
The scraping engine now automatically identifies domains that only require background verification checks and skips unnecessary behavioral wait times. This cuts request latency on these sites by up to 4 seconds.
Automatic IP block recovery
The platform now automatically detects connection-level proxy blocks and triggers a fresh egress route rotation. Instead of failing immediately when an IP is flagged, your requests are seamlessly retried with a clean IP.
Enterprise detection compatibility
We have added automatic detection and optimized routing pathways for multiple advanced, enterprise-grade bot protection systems. This ensures significantly higher success rates when scraping highly secured enterprise and retail websites.
Direct deep-links for support ratings
Support ticket resolution emails now include direct deep-links that automatically highlight and focus the satisfaction rating widget on our platform. This lets you leave feedback on your support experiences in fewer clicks.
New anti-bot benchmarks
You can now view live difficulty scores across all our comparison pages. This helps you compare real-world performance against modern detection systems to make informed integration decisions.
Improvements
8Parallel multi-engine requests
Optimized request routing to initiate parallel connection strategies from the very start of a request. This removes sequential delays on standard request paths, leading to significantly faster overall response times.
Budget-aware request timeouts
Request connection timeouts are now dynamically adapted to match your remaining request time budget rather than using a static default. This ensures requests either complete successfully or fail-fast within your specified limits.
Sticky session documentation updates
Updated our developer documentation and AI-agent reference files to include details on the 10-minute automatic exit IP rotation cycle. This provides clearer guidance on maintaining long-lived scraping sessions.
Improved session persistence reliability
Fixed a routing key mapping issue during back-end migrations to ensure persistent scraping sessions remain uninterrupted. Your sticky sessions and designated identities will transition seamlessly without connection drops.
General bug fixes and improvements
Plus 34 internal improvements for better reliability and performance.
Optimized request latency
We streamlined our pre-scrape routing pipeline to eliminate redundant checks. This reduces initial routing latency by over two seconds for standard requests, allowing your scraping tasks to start much faster.
Consistent large-scale exports
We added automatic lock renewals during large-scale data export cycles. This prevents tasks from expiring mid-cycle, eliminating duplicate files and ensuring consistent data delivery.
Accessible comparison metrics
We added full ARIA progressbar semantics to the difficulty scorebars on our comparison pages. This ensures screen readers and assistive devices can easily interpret our real-world benchmark data.
Bug Fixes
9Enhanced verification polling stability
Replaced periodic polling checks with event-driven triggers when waiting for verification challenges to resolve. This eliminates unnecessary delays and ensures requests proceed the millisecond a verification check is solved.
Prevented early challenge timeouts
Fixed a race condition where a pre-existing stale-content timeout could trigger a fraction of a second before the newer rejection grace period expired. This prevents premature failures on slow-loading verification pages.
Accurate billing for failed pages
Corrected a routing and billing issue on Adidas homepage URLs where queries with empty parameters returned a bad request error disguised as a successful HTTP 200. These requests are now properly handled so you only pay for successful scrapes.
Seamless interactive challenge recovery
Fixed an issue where transient background verification failures would permanently cancel subsequent verification attempts. The system now robustly continues the full verification flow even if initial background checks are temporarily rejected.
Targeted content wait on retry
Solved an issue where retried requests that successfully bypassed verification pages would prematurely skip waiting for target selectors. The engine now correctly waits for your specified elements to render post-bypass.
Keyboard focus collision fix
Resolved a UI focus issue where simultaneously mounted mobile and desktop versions of the customer support feedback widget would compete for keyboard focus. This improves accessibility and usability on responsive layouts.
Challenged request recovery
We enabled advanced request properties when encountering temporary verification screens. This automatically rotates IPs with full browser characteristics, boosting success rates on guarded domains.
Batch poller reliability
We resolved an issue where batch workers performed redundant retries after losing their execution locks. This prevents duplicate tasks and ensures much more stable long-running operations.
Resilient batch exports
We added error isolation to the export backfill process. A transient error with a single record will no longer abort the entire run, ensuring your successful scraped data is safely delivered.
Plus 47 internal changes for stability and performance.