xHamster 2.6.5

xHamster 2.6.5

This release fixes problems found while testing public video, listing and download workflows.

• Restore shorts and moments extraction from current page data and DOM layouts.
• Keep profile shorts on the correct listing route and preserve regional URL hosts.
• Read listing metadata from hydrated page data so later cards retain thumbnails and durations.
• Avoid incompatible quality filters for VR search and recognize Full HD and UHD videos correctly.
• Handle missing short metadata safely and add compatible user, video and shorts aliases.
• Route shorts and moments through the CLI.

The package now uses eaf-base-api>=4.2.5 from PyPI. No local source checkout is required.

unofficial-api-for-xhamster 2.6.3

Release 2.6.3 fixes provider URL handling and pagination.

Changes:

• Preserve profile tabs and filters during pagination, replacing an existing numeric page suffix.
• Use the resolved localized host for listing requests and relative video links.
• Replace account playlist page query keys without duplication.
• Require eaf-base-api>=4.2.2 for shared URL composition and resolved-response metadata.
• Add focused offline URL regression tests; verify the repaired behavior with bounded live metadata/listing checks.

Install from PyPI:

uv pip install unofficial-api-for-xhamster==2.6.3

Source commit: d3210cc

2.6.2

Changelog

• Fixed an issue where invalid login credentials would report a successful login

2.6.1

What's Changed

Improvements & Resilience

• Layout Anchor Validations: Added structural anchor checks to detect upstream markup shifts.
• Metadata Fallbacks: Resilient fallback selectors for video and pornstar metadata.
• Stream URL Resolution: Enhanced stream URL parsing and error reporting.
• Unit Tests: Added snippet-based unit tests for video and pornstar extraction.

Core Engine Upgrade

• Bumped dependency to eaf-base-api>=4.2.0.
• Adopted fetch_content and @download_errors.
• Download failures now raise DownloadFailed; cancellation raises DownloadCancelled.

2.6

Changelog

• Added a lot of new video attributes. See docs :)

2.5

Changelog

• See: https://docs.echteralsfake.me/xhamster/#changelog

2.4

Changelog

I am too lazy to write a dedicated changelog for all my APIs now, so here's a summary

• Raised Python version to 3.12 (because I use Python 3.12 specific async logic + type hints)

• Improved memory usage by 99%

• Dataclass models for all objects with slots

• Added new exceptions that actually work now

• Improved parsing speed for video iterators a lot because we can actually just scrape the data from the page directly and fill it into the video object. Just read the docs lol

• I have added an option to the iterators, so you can actually say if the videos should be in the correct order or not which has an influence on scraping speed

• You can now pass custom on_error functions into the iterators and handle video re-trying yourself

Oh yeah..

Switched to AGPL

Does this break my existing code?

Yes.

2.2

Changelog

• raised python version

2.1

• fixed a bug

Version 2.0

Changelog

Switching to Async

Warning

This release absolutely breaks your existing code. Read the documentation (once I updated it)

• Added more stuff to the Short class
• Replaced httpx with curl_cffi for HTTP requests, enabling browser impersonation ( chrome ) to bypass
  bot protections
• Converted all network methods to async/await ( fetch , download , legacy_download , get_segments ,
  get_m3u8_by_quality , list_available_qualities , download_segment )
• Converted Helper.iterator from a synchronous ThreadPoolExecutor -based generator to an async generator
  using asyncio.Task
• Replaced ThreadPoolExecutor in the HLS segment downloader with asyncio.Semaphore + asyncio.
  as_completed
• Added async_generator_to_sync bridge utility for synchronous iteration
• Added fast concurrent multipart range downloading to legacy_download ( allow_multipart=True by default)
• Added ios_support parameter to download for stricter audio codec filtering during remux
• Fixed progress bars to handle unknown/zero total sizes gracefully instead of dividing by zero
• Fixed multipart downloads failing on compressed streams by requesting Accept-Encoding: identity for range
  requests
• Changed remux progress tracking from packet count to byte-based for more accurate reporting