Faster Scraping and Programmatic Triage
This release introduces secure cross-account S3 deliveries via AWS IAM roles, eliminating the need to manage static credentials. We have also optimized our scraping tiers for low-volume domains and short responses to reduce credit usage and improve success rates. Additionally, you can now reliably retrieve hidden pricing for MAP-restricted products on Amazon.
Deferred requests now resume with millisecond-level precision instead of waiting a flat 5-second ceiling. This significantly increases your overall scraping throughput.
Proxy routing now distributes concurrent requests across a wider pool of exit IPs. This prevents traffic concentration on single IPs and dramatically reduces rate-limiting blocks.
The platform now prevents unnecessary escalation to expensive browser-based scraping tiers during transient network or proxy degradations. This keeps your API usage highly cost-effective.
New Features
3Resolve tickets programmatically
You can now resolve or reopen support tickets directly using your API keys, removing the need for user JWTs. This simplifies the process of automating ticket lifecycles within your own systems.
Automate ticket triage classifications
You can now programmatically update ticket priorities and tags using API key authentication. This lets you integrate automated ticket classification systems directly into your workflow.
Merge duplicate support tickets
We have introduced a new API endpoint to merge duplicate support tickets programmatically by reference. This allows your team to automate the cleanup of redundant incoming customer requests.
Improvements
12Millisecond-level rate limit queueing
Deferred requests now resume with millisecond-level precision instead of waiting a flat 5-second ceiling. This significantly increases your overall scraping throughput.
Concurrent proxy IP distribution
Proxy routing now distributes concurrent requests across a wider pool of exit IPs. This prevents traffic concentration on single IPs and dramatically reduces rate-limiting blocks.
Predictive scraper tier escalation
The platform now prevents unnecessary escalation to expensive browser-based scraping tiers during transient network or proxy degradations. This keeps your API usage highly cost-effective.
Humanized request pacing delay
Scraping requests now incorporate natural inter-request delays and paced retries on failure. This human-like behavior reduces detection rates on websites with strict bot protection.
AI agent documentation endpoints
We have introduced new agent-friendly documentation routes. This allows developers' AI assistants to easily parse and query our API schemas for seamless integration.
Enforced rotating proxy blacklists
IP blacklists are now strictly enforced in rotating proxy mode. The system automatically re-rolls to a clean connection if a known bad or burned exit IP is detected during a request.
Automatic JSON response parsing
When targeting JSON API endpoints, the parsed response body is now cleanly exposed in the content JSON field. This makes extracting structured data much more straightforward.
General bug fixes and improvements
Plus 47 internal improvements for better reliability and performance.
Hidden Amazon Price Retrieval
Automatically extract prices for products under Amazon's Minimum Advertised Price restrictions. The API now retrieves hidden prices instead of returning empty fields.
Cost-Optimized Low-Volume Scraping
Low-volume domains now transition to cheaper scraping tiers much faster once success is proven. This prevents requests from getting stuck on expensive browser-rendering tiers, saving you credits.
Faster Browser Escalation Retries
Requests requiring automatic browser escalation now retry on an optimized schedule instead of sharing standard delays. This significantly reduces overall latency and ensures jobs complete faster.
More Responsive Request Dispatching
Eliminated unexpected latency spikes when starting requests after the scraper has been idle. Queue poll delays are now strictly bounded, resulting in faster and more consistent response times.
Bug Fixes
7Shielded active scraping sessions
Active scraping sessions are now protected from premature background cleanup tasks. This eliminates unexpected session terminations and ensures long-running requests finish successfully.
Empty HTML content validation
Requests returning empty HTML shells with no visible content are now correctly identified and escalated. This ensures you always receive complete and valid page data.
Auto-recharge invoice download fix
We resolved a dashboard bug where auto-recharge transactions were displayed as non-clickable items. You can now easily view and download invoice receipts for all transactions.
Clean ticket reply threads
Submitting whitespace-only notes will no longer generate empty, confusing ticket responses in the system. This keeps your support communication clean and readable.
Robust S3 Credential Validation
We added validation to S3 delivery configurations during creation and credential rotation. This prevents configuration errors, such as missing external IDs, before they cause delivery failures.
Improved Handling of Short Responses
Restored reliable escalation to browser rendering for pages that return extremely short, non-JSON initial responses. This ensures that minimal landing pages are properly rendered and resolved.
Reliable LLM Configuration Setup
Resolved an initialization issue that could prevent the automatic configuration of LLM-based extraction models. Your AI-assisted parsing tasks will now configure and run reliably on startup.
Plus 61 internal changes for stability and performance.