Name | Modified | Size | Downloads / Week |
---|---|---|---|
Parent folder | |||
langroid-0.58.0-py3-none-any.whl | 2025-07-10 | 414.1 kB | |
langroid-0.58.0.tar.gz | 2025-07-10 | 364.8 kB | |
0.58.0 source code.tar.gz | 2025-07-10 | 57.5 MB | |
0.58.0 source code.zip | 2025-07-10 | 58.0 MB | |
README.md | 2025-07-10 | 1.4 kB | |
Totals: 5 Items | 116.3 MB | 0 |
🕷️ Crawl4AI Integration
(Thank you @abab-dev !)
This release introduces Crawl4AI as a powerful new web crawling option in Langroid.
Highlights
- Browser-based crawling with Playwright - handles JavaScript-heavy sites and SPAs
- No API key required - runs entirely locally
- Multiple extraction strategies - CSS selectors, LLM-based, regex patterns
- Deep crawling support - recursively crawl entire websites
- Advanced content filtering - remove ads, sidebars, and irrelevant content
Quick Start
:::bash
pip install "langroid[crawl-4-ai]"
crawl4ai setup # One-time Playwright download
:::python
from langroid.parsing.url_loader import URLLoader, Crawl4aiConfig
loader = URLLoader(urls=["https://example.com"], crawler_config=Crawl4aiConfig())
docs = loader.load()
Documentation
Other Improvements
- CLI framework migration from Typer to Fire in
chat_search.py
- Enhanced URL loader framework with better crawler factory pattern
- Added comprehensive tests with mocking for CI compatibility
Full Changelog: https://github.com/langroid/langroid/compare/v0.57.0...v0.58.0