More Reliable Scraping and Accurate Billing
This release focuses on dramatically improving job stability, recovery, and billing accuracy during transient network disruptions or target rate limits. We've introduced smarter recovery for multi-stage requests, prevented duplicate job runs, and patched a critical vulnerability in our connection stream handler. Additionally, browser-based rendering is now more resilient with automatic startup retries and self-healing extension caches.
We improved our billing reconciliation process to handle ambiguous API transactions during background failovers. This guarantees accurate usage tracking even during transient network issues.
The API now automatically filters out completed jobs during background recovery events. This prevents duplicate webhook deliveries and redundant scraping attempts, saving your API credits.
Scraping jobs requiring multiple stages are now fully preserved during recovery operations. This prevents active requests from being prematurely marked as timed out or failed.
Improvements
4Robust billing and settlement recovery
We improved our billing reconciliation process to handle ambiguous API transactions during background failovers. This guarantees accurate usage tracking even during transient network issues.
Accurate refunds for terminated jobs
API refunds and terminal settlements are now serialized to eliminate rare race conditions. This ensures refunds for failed requests are applied promptly and accurately to your account.
Reliable handling of paced requests
Jobs delayed due to target-site rate limits now remain fully active and recoverable. This prevents requests from timing out prematurely on sites requiring polite, low-concurrency pacing.
General bug fixes and improvements
Plus 14 internal improvements for better reliability and performance.
Bug Fixes
4Preventing duplicate job execution
The API now automatically filters out completed jobs during background recovery events. This prevents duplicate webhook deliveries and redundant scraping attempts, saving your API credits.
Robust multi-stage job preservation
Scraping jobs requiring multiple stages are now fully preserved during recovery operations. This prevents active requests from being prematurely marked as timed out or failed.
Reliable browser initialization on retries
We resolved an issue where incomplete extension downloads caused scraping retries to fail. The platform now automatically clears corrupted cache directories to ensure smooth page loads.
Improved browser engine discovery
We added retry logic with backoff for the browser rendering engine's initialization and discovery. This eliminates transient startup errors when launching browser-based scraping jobs.
Security
1Resolved connection stream vulnerability
We patched a vulnerability (CVE-2026-63374) in our connection streaming components to keep all API communication secure. This update prevents potential service disruptions.
Plus 14 internal changes for stability and performance.