Page in Words: the web, read out as plain words. No key, no sign-up. Put the address after the slash, or in url=. The answer is the page as Markdown: https://www.pageinwords.com/read/https://example.com/some/article https://www.pageinwords.com/read?url=https://example.com/some/article From a shell: curl https://www.pageinwords.com/read/https://example.com/some/article curl -G https://www.pageinwords.com/read \ --data-urlencode 'url=https://example.com/page?a=1&b=2' Other formats: text without the marks, cleaned-up HTML, or JSON with the title, the final address, the status and the content: https://www.pageinwords.com/read?format=text&url=https://example.com/ https://www.pageinwords.com/read?format=html&url=https://example.com/ https://www.pageinwords.com/read?format=json&url=https://example.com/ curl -H 'Accept: application/json' \ https://www.pageinwords.com/read/https://example.com/ Put url last, or percent-encode it: an address written out as it is runs to the end of the line, with its own ? and & in it. Redirects are followed and X-Source-Url says where the page came from. Images are left out; images=1 keeps them as links. A long page is cut at 400 kB and X-Incomplete says so. HTML, plain text, Markdown, JSON and XML are read; PDFs and other files are not.