All releases
v3.9.2
June 27, 2026
Patch
In this release
New2
Improved6
Fixed6
Security2
Internal38
Total54

Smarter Testing, Failovers, and Enhanced Security

This release introduces a seamless transition from our homepage demo to the interactive developer playground, alongside blog-embedded sandbox links for instant code testing. We have also hardened our security architecture with guaranteed tenant isolation and improved billing accuracy with automated search credit refunds. Finally, robust LLM failovers and enhanced CAPTCHA handling ensure higher success rates across all scraping tasks.

Highlights
Homepage Demo Integration

Test URLs on our homepage and instantly open your extraction results in the playground with a single click. This lets you inspect full results, experiment with configurations, and export code without restarting your scrape.

Automated Search Credit Refunds

Guarantees billing integrity for batch search engine queries by automatically refunding credits for unsuccessful, expired, or failed tasks. You will only be billed for successfully completed search jobs.

Guaranteed Tenant Isolation

Introduces atomic slot-claiming in our session pools to ensure absolute user isolation. This guarantees concurrent users never share browser profiles, active cookies, or session histories.

New Features

2

Homepage Demo Integration

Test URLs on our homepage and instantly open your extraction results in the playground with a single click. This lets you inspect full results, experiment with configurations, and export code without restarting your scrape.

Blog Sandbox Integration

One-click sandbox testing is now built directly into our technical guides. You can launch blog post code examples straight into the interactive developer playground, pre-configured and ready to run.

Improvements

6

General improvements

This release includes general improvements and maintenance updates.

General bug fixes and improvements

Plus 9 internal improvements for better reliability and performance.

Type-Safe Node SDK Search

The Node SDK now includes full TypeScript support for the engine selection parameter in both single and batch search requests. Developers can now build programmatic queries with complete type safety and editor autocomplete.

Flexible LLM Failover

Our AI-assisted extraction pipeline now supports robust fallback model routing. If your primary AI model encounters rate limits, the system seamlessly cascades to your fallback model to prevent job interruptions.

Smarter Pricing Interface

We have simplified our pricing page to highlight our Smart Router benefits and clear up plan selection. A streamlined navigation experience makes it easier to evaluate options and find the perfect tier.

Smarter Webhook Balance Protection

Your API credits are now protected from invalid webhooks. Balance debit logic runs after webhook destination validation, ensuring credits are never charged if a webhook URL fails initial checks.

Bug Fixes

6

decrypt-secrets.sh pipefail kills script on missing grep keys

- Add to all grep pipelines in feature-flag validation and monitoring config sections.

Automated Search Credit Refunds

Guarantees billing integrity for batch search engine queries by automatically refunding credits for unsuccessful, expired, or failed tasks. You will only be billed for successfully completed search jobs.

Reliable Session Rotation

Ensures correct session and IP rotation triggers are reached during health checks. This guarantees consistent evasion of detection systems and significantly higher success rates for long-running scraper tasks.

Parallel CAPTCHA Solving

Resolves a timeout conflict where CAPTCHA solving challenges were prematurely cancelled during parallel dispatches. CAPTCHA routines now have sufficient time to complete, boosting success rates on protected pages.

Robust Job Routing

Fixes a queue routing error where search engine query jobs could be processed by generic workers lacking the required session context. Jobs are now correctly pinned to matching worker queues.

General Bug Fixes

Includes several stability improvements across our platform. We resolved unreachable administrative endpoints, fixed JWT authentication errors, and corrected playground cost calculation inaccuracies.

Security

2

Guaranteed Tenant Isolation

Introduces atomic slot-claiming in our session pools to ensure absolute user isolation. This guarantees concurrent users never share browser profiles, active cookies, or session histories.

Enhanced SSRF Defenses

Hardens internal network security with robust Server-Side Request Forgery validation. Additional checks protect custom storage configurations and background webhook tasks from unauthorized requests.

Plus 38 internal changes for stability and performance.