mirror of
https://github.com/unclecode/crawl4ai.git
synced 2026-06-11 00:08:01 +00:00
158231be60
Detect single full-viewport <frameset> pages and automatically navigate
to the child frame's URL before the crawl pipeline runs, so JS execution,
overlay removal, screenshots, and HTML capture all operate on real content
instead of the empty frameset shell. Fixes false-positive antibot detection
on frameset wrapper sites (e.g. variosolar.de → varioself.de).
- Add follow_frames param to CrawlerRunConfig (default True)
- Add _detect_frameset_redirect() to AsyncPlaywrightCrawlerStrategy
- Insert Phase 0 frameset detection before Phase 1 (pre-wait JS)
- Update docs: parameters.md, content-selection.md, page-interaction.md,
complete-sdk-reference.md