# Web articles (/docs/supported-content/web-articles)



The general article extractor is designed for public editorial pages with a clear title and main
body in the initial HTML response.

## Usually supported [#usually-supported]

* News, magazine, blog, and documentation articles with semantic text.
* Headings, paragraphs, lists, quotations, and useful links.
* Public images that can be downloaded safely and converted for EPUB use.
* Canonical HTTP or HTTPS article URLs.

## Common limitations [#common-limitations]

* Pages whose article appears only after client-side JavaScript runs.
* Login walls, paywalls, consent walls, and bot challenges.
* Private, deleted, region-restricted, or session-bound pages.
* Homepages, search results, application dashboards, and other pages without a single main article.
* Extremely large or structurally complex documents.

The extractor needs a meaningful readable body. Navigation, scripts, forms, tracking elements,
advertisements, and unrelated page chrome are removed.

## Images and interactive content [#images-and-interactive-content]

Images are included on a best-effort basis. Remote images are constrained by count, size, response
time, and supported formats; a missing image does not necessarily make the article fail. WebP and AVIF
images can be converted for broader eReader compatibility.

Interactive embeds and JavaScript do not run inside an EPUB. Mermaid diagrams are represented by a
text fallback rather than rendered interactively.

## Public access test [#public-access-test]

Open the canonical article URL in a signed-out private browser window. If the article body is not
available there without interaction, the server is unlikely to be able to capture it.

For failures, see [Article capture troubleshooting](/docs/troubleshooting/article-capture).