Summary of an article published on The Web Scraping Club on 2026-02-15 (free to read). Read the full article
Scraping means retrieving and using someone else’s data, so the copyright question comes up sooner or later. Public availability is not ownership. Articles, images, audio, video and sometimes code are protected, and even a small part of an article is not a safe harbor. Descriptions that look like plain data fields can also be creative work, and training LLMs on scraped pages is still an open legal area.
We go through the situations where infringement tends to happen, from copying content and media to scraping for LLM training. Then we cover the guidelines we follow: scrape facts rather than expression, transform instead of replicate, and avoid storing raw pages. We also describe two standardized processes, a copyright risk check before adding any field and a scraper built so it is hard to do something dumb. We close with examples of safe-ish and trouble-ish setups, and with the risks that have nothing to do with copyright.
What’s inside
- Why images get their own “danger zone” treatment and which safer routes exist for them
- The questions to ask before adding a field or selector to a scraper, including whether your output would compete with the source
- Five use cases (price tracker, catalog clone, news aggregation, review analytics, business directory) ranked from safe-ish to high risk
- Non-copyright traps: DMCA Section 1201 circumvention, robots.txt, terms of service and scraping behind a login
In the full article: The full list of risk-check questions, the pipeline defaults for what to keep and what to discard, and the detailed reasoning behind each use case example.