All releases
v2.49.0
May 18, 2026
Minor release
In this release
New5
Improved9
Fixed15
Security2
Internal22
Total53

Bug Fixes & Stability

This release includes 11 updates.

Highlights
Expanded HTTP Method Support

You can now use any HTTP method and include custom request bodies in your scraping tasks. This provides the flexibility needed to interact with complex web applications and authenticated endpoints.

New Free Developer Tool Suite

Access 12 new micro-tools directly on our platform, including a JSON validator, HTTP header checker, and robots.txt tester. These utilities help you debug and verify web requests without leaving your browser.

Worker Performance Optimization

We have optimized our data extraction engine to significantly reduce latency and unblock processing loops. These changes ensure faster response times and higher throughput for your most demanding scraping jobs.

New Features

5

Expanded HTTP Method Support

You can now use any HTTP method and include custom request bodies in your scraping tasks. This provides the flexibility needed to interact with complex web applications and authenticated endpoints.

New Free Developer Tool Suite

Access 12 new micro-tools directly on our platform, including a JSON validator, HTTP header checker, and robots.txt tester. These utilities help you debug and verify web requests without leaving your browser.

add proxy country selector to scrape playground

- Adds a country dropdown in the playground Network section (visible when proxy rotation is enabled).

add http_method column to usage_tracking for abuse detection

Add a dedicated CODE column to CODE so operators can query mutation volume (DELETE/PUT/PATCH usage) by user for abuse detection.

implement BFS link extraction for map endpoint — respect max_depth

The CODE endpoint accepted CODE (0–10) but only ever fetched depth-1 links regardless of the value.

Improvements

9

Worker Performance Optimization

We have optimized our data extraction engine to significantly reduce latency and unblock processing loops. These changes ensure faster response times and higher throughput for your most demanding scraping jobs.

Support for Large Request Payloads

We have removed system limits on request body sizes, allowing you to send payloads larger than 2MB. This enables more complex interactions with enterprise sites that require large data transfers.

Optimized Webhook Delivery

A new concurrency management system ensures that one slow destination endpoint won't delay deliveries to your other webhooks. This provides more predictable and reliable event streaming at high volumes.

Accurate System Health Monitoring

The health check endpoint now provides a real-time status of critical system dependencies rather than a static response. This allows your monitoring tools to receive a more accurate picture of platform availability.

Graceful Job Completion

Improved system shutdown procedures ensure that in-flight scraping jobs and webhooks are given dedicated time to finish gracefully during maintenance. This reduces the risk of interrupted data delivery.

General bug fixes and improvements

Plus 5 internal improvements for better reliability and performance.

Increased browser scraping capacity

We have significantly scaled our backend infrastructure to handle heavy browser-based scraping workloads. You will experience drastically fewer pool exhaustion errors and improved success rates during high-concurrency operations.

Automatic search query retries

Search API requests now feature intelligent exponential backoff when encountering upstream rate limits. This minimizes failed requests and ensures smoother data retrieval without requiring manual retry logic on your end.

Comprehensive API documentation updates

We have expanded our developer documentation to cover essential topics like authenticated scraping, webhooks, and account limits. This includes updated guides for our SDKs and custom proxy configurations to help you build faster.

Bug Fixes

15

Improved Geolocation Logic

Your location settings are now correctly applied across all request types, including browser-based scraping. This ensures you receive localized content reliably without being blocked by regional filters.

Reliable Domain Scraping

Fixed routing issues affecting major retail and Q&A platforms that previously caused intermittent failures. These domains now correctly utilize the necessary resources for successful and consistent data extraction.

General Platform Stability

This update resolves several minor issues, including support file upload errors, dashboard filter bugs, and broken shortcut links to the browser extension.

Reliable POST request bodies

POST requests sent through our scraping API now correctly pass the request body payload to the target server. A routing bug that caused payloads to be silently dropped under specific configurations has been fully resolved.

Accurate batch rate limiting

Batch endpoints for scraping and searching now correctly apply rate limits based on the total number of items in the request. This ensures fairer platform resource distribution and consistent performance for all API consumers.

Accurate request failure analytics

Analytics dashboards now correctly categorize rejected requests and handle malformed HTTP methods gracefully. This provides much clearer insights into request failures and ensures your usage metrics load without interruption.

Corrected custom proxy parameters

Custom proxy reordering parameters are now accurately validated as query parameters rather than request body fields. This prevents unexpected validation errors when configuring complex proxy sequences.

enforce 25 MB per-ticket attachment quota in create_ticket

- Import CODE from CODE into CODE.

use FlexibleUser on attachment upload/download endpoints

CODE and CODE in CODE used CODE which only accepts X-API-Key or CODE tokens.

file uploads always fail on /dashboard/support/new

- Upload endpoint was creating a CODE DB row with both CODE and CODE NULL, violating the CODE CHECK constraint and returning 500 on every upload.

MissingGreenlet on /api/v1/usage/estimate — capture user_id eagerly

- Fixes 500 error on CODE and CODE caused by CODE.

correct misleading NULL comment for http_method DEFAULT backfill

Fixes inaccurate documentation in migration 0140 and the UsageTracking model.

neutralize liability language in HTTP method relay docs

Replace encouraging/directive language about DELETE relay ("Use DELETE to relay REST API delete operations") with neutral, permission-caveated language across SDK docs and versioned OpenAPI spec.

strip PUT/PATCH/DELETE from scrape API and playground UI

- Narrows allowed HTTP methods from GET/POST/PUT/PATCH/DELETE to **GET/POST only** in both the Pydantic schema (returns 422 for disallowed methods) and the playground UI dropdown.

remove dead budget check superseded by pre-LLM guard

Removes the duplicate budget check block at lines 507-519 in CODE that became dead code after added an earlier pre-LLM guard at lines 480-488.

Security

2

Secure SSL Validation Utility

Enhanced the security of our SSL checking tool by implementing stricter network access controls. These updates prevent unauthorized network probes while maintaining accurate certificate validation for users.

Enhanced SSL checker security

We addressed a vulnerability in our API's SSL checker tool that could have allowed restricted local network addresses to be queried. The tool now strictly filters out all internal IP ranges to ensure secure external operation.

Plus 22 internal changes for stability and performance.