Extract links, a table or the article

Pulls a known shape out of a page: its links, images, first table, meta tags, structured data, readable article or the article as Markdown.

Extract the kind from the page

Details #

kind is:

links: a list of {text, href}.
images: {src, alt}.
table: a list of dictionaries from the first table (or the one at selector), keyed by its header cells.
meta: description, Open Graph tags, the canonical address.
jsonld: the page's structured data.
article: {title, byline, text}, the reader-mode text.
markdown: the article as Markdown.
feed: the page's RSS, Atom or JSON feed as [{title, url, date, author, summary}], found through <link rel="alternate">, or the page itself when it is one.
list: the page's main repeating list as rows, the same as Scrape a page into rows with no selector.

selector limits it to one part of the page.

Parameter	Type	Required	Values	What it does
`kind`	one of	Yes	`links` `images` `table` `meta` `jsonld` `article` `markdown` `feed` `list`	Required.
`browser`	one of		`safari` `chrome` `front`
`html`	text
`impersonate`	one of		`auto` `chrome` `safari` `none`
`into`	text			The variable that receives the result.
`sample`	any value
`selector`	text			Only this part.
`signIn`	text
`tab`	any value
`timeout`	number			, tab.
`url`	text			, html, via, signIn, impersonate: Read a page nobody has open: `url` is fetched the way a browser would (`via: fetch`, the default) or opened in a background tab (`via: tab`); `html` is a string such as `{{ res.body }}`.
`via`	one of		`auto` `extension` `appleScript` `fetch` `tab`

Every step also accepts label, onFailure, timeout and dryRunSafe.

Copy the article as Markdown

Open in Switchboard to try it

The article is a good guess, not a promise: a page with no clear article may give the largest block of text.