All releases
v3.12.0
July 2, 2026
Minor release
In this release
New1
Improved8
Fixed8
Security2
Internal39
Total58

Resilient Scraping and Smarter Retries

This release introduces improved authentication compatibility for automated workflow tools, ensuring seamless usage reporting. We have also resolved playground visualization issues and standardized billing document downloads for a more reliable integration experience.

Highlights
Unlocked playground preview

You can now view the unblurred raw output of your first scrape in the sandbox playground. This lets you inspect the precision and structure of the returned data instantly before signing up.

Automatic routing optimization

The system now handles rate limits with an automatic backoff instead of immediate escalation to browser-based tiers. This preserves your high-performance credits and reduces billing costs.

Resilient transient network recovery

Temporary network disruptions, such as connection resets, are now properly retried instead of being treated as permanent failures. This greatly increases success rates for large-scale jobs.

New Features

1

Unlocked playground preview

You can now view the unblurred raw output of your first scrape in the sandbox playground. This lets you inspect the precision and structure of the returned data instantly before signing up.

Improvements

8

Automatic routing optimization

The system now handles rate limits with an automatic backoff instead of immediate escalation to browser-based tiers. This preserves your high-performance credits and reduces billing costs.

Intelligent browser-rendered retries

The browser-rendering tier now automatically rotates proxies and retries on transient errors or timeouts. This significantly increases success rates on highly protected sites.

First-party invoice delivery

Billing emails now deliver invoice documents as secure, first-party static PDFs rather than dynamic third-party views. This ensures documents load instantly inside all secure email clients.

Optimized browser-free routes

Our routing engine now skips browser-specific checks entirely when bypassing rendering execution. This reduces handshaking overhead and accelerates response times for static-content scrapes.

Enhanced extraction reliability

Our AI selector and schema generation engine is now hardened against premature output truncation. It parses complex page layouts with nested formatting robustly, while protecting output integrity.

General bug fixes and improvements

Plus 32 internal improvements for better reliability and performance.

Consistent detection priority handling

Our request evaluation prioritizes block screens over interactive challenges from detection systems. This ensures consistent handling and highly accurate error reporting during scraping tasks.

Robust session diagnostic logging

We have introduced detailed background logging for session baseline failures. This improves troubleshooting for edge-case connection drops and helps resolve scraping issues much faster.

Bug Fixes

8

Resilient transient network recovery

Temporary network disruptions, such as connection resets, are now properly retried instead of being treated as permanent failures. This greatly increases success rates for large-scale jobs.

Rate-limit backoff pacing

The queue now respects backoff delays before retrying rate-limited requests instead of re-enqueuing them immediately. This maximizes success rates and avoids wasting your retry budget.

Accurate block detection priority

The request engine now prioritizes block page checks before launching challenge solvers. This prevents the scraper from misclassifying block pages and ensures smarter automated recovery.

Smarter timeout budget tracking

Browser engine retries now strictly respect your remaining timeout budget. The engine will fail-fast if a crash happens near your limit, returning responses sooner instead of starting a new session.

Unified registration requirements

We aligned client and backend password validation rules. The signup and settings forms now correctly enforce the 10-character minimum (with a letter and digit), preventing unexpected errors.

Seamless usage tracking authentication

You can now query your API usage using standard API keys instead of bearer tokens. This resolves authentication issues during integration tests on external automation platforms.

Accurate success tracking in Playground

The interactive sandbox now correctly reveals scraped content on your first successful attempt, even if earlier attempts failed. This ensures successful previews are immediately visible.

Reliable billing document downloads

Billing invoice downloads are now pinned to the correct document type so your browser can open them immediately. This prevents files from being misidentified or downloaded with incorrect formats.

Security

2

Secure payment document links

We added strict protocol checks to verify all payment invoice references over secure HTTPS connections before retrieval. This enhances transactional safety and prevents unencrypted fetches.

Strict inline script security

We updated all inline scripts on our homepage to enforce per-request cryptographic nonces, aligning fully with strict Content Security Policy guidelines to prevent cross-site execution.

Plus 39 internal changes for stability and performance.