83,999 stars · BSD-3-Clause, read from /blob/main/LICENSE on 2026-09-27; no restriction on commercial use, no enterprise carve-out · v0.4.15 (2026-08-23), read from /releases/latest and confirmed on PyPI · Track this in Scout
A Python scraping toolkit that stores structural context alongside each selector, so an element can still be found after a site is redesigned.
▶Repo detailsthe review · specs · pros & cons · install
What it is
Scrapling is a scraping framework for the Python programming language, covering everything from a single page fetch to a large crawl across a whole site. It offers three ways to fetch a page — a plain request, a "stealthy" mode aimed at anti-bot systems, and a real browser — behind one way of reading the result.
What it is good for. Anyone who runs a scraper that has to keep working for months rather than a weekend. Price monitoring, job listings, competitor pages and public records all break the same way: the site is redesigned and the selector (the short instruction that says which part of the page you want) quietly matches nothing. Scrapling stores structural context alongside each selector so the element can be relocated instead of silently returning empty. It also packages a command-line shell for pulling data out without writing a program, and an MCP server (a standard way to let an AI assistant drive a tool) so a model can run the scraping itself.
- The adaptive relocation of elements attacks the single biggest running cost of scraping, which is repairing selectors. None of the alternatives below offers it.
- BSD-3-Clause, with no conditions on commercial use, so it can sit inside a paid product.
- Genuinely fast-moving. Four releases between 26 July and 23 August 2026, and code landed on 26 September 2026.
- Installing it has two stages and is easy to get wrong.
pip install scraplingon its own gives you only the page reader, and the first line that fetches anything then fails with a missing-module error. You have to install the extra and runscrapling installseparately. - Neither the README nor the documentation states any memory or disk figure anywhere. The install step downloads browsers, and the download size is not published, so you cannot plan disk space. Stealth fetching runs real browsers, which is heavy in a way the documentation never quantifies.
- Still before version 1.0, at v0.4.15. One release, v0.4.14, exists only to work around another installer refusing v0.4.13. The way you call it can change between versions.
- scrapy/scrapy
64,264 stars, the long-established framework for the same job; far more proven and far more to learn, with no adaptive selectors.
Track this in Scout 
unclecode/crawl4ai84,283 stars, covered in Edition 25; aimed at turning pages into clean text for a language model rather than at extracting specific fields.
Track this in Scout
apify/crawlee25,911 stars, covered in Edition 25; manages the queue and the retries very well, and is JavaScript rather than Python.
Track this in Scout
python3 -m venv venv source venv/bin/activate # on Windows: venv\Scripts\activate pip install "scrapling[fetchers]" scrapling install
