Two paragraphs from a colleague, forty kilobytes of markup, and a CMS that now renders them in Calibri for reasons nobody can explain. Paste it here and get back the HTML you thought you were pasting.
Word was never trying to write HTML for a website. It was trying to write HTML that Word itself could open again later and find every setting exactly where it left it: the fonts, the spacing, the language of each individual span, the list numbering, the lot. All of that goes into the file as mso- properties and namespaced tags nobody outside Redmond has ever needed to read.
Which is fine, right up until the moment that file becomes the source of a page on your website. Then it is not a document any more, it is a paragraph of text wearing a forty kilobyte coat.
The first is lists. Word does not write lists, it writes ordinary paragraphs with a bullet character glued on the front and a note in the style attribute explaining what it meant. Strip the styles without reading that note and you get a run of paragraphs starting with a middle dot, which looks almost right and behaves nothing like a list. Those are rebuilt here, nesting and numbering included.
The second is Google Docs, where bold and italic are inline styles and nothing else. Remove the styles first and you have silently deleted every emphasis in the document. So they are promoted to real <strong> and <em> tags before anything is thrown away.
Both work, and you do not have to say which you are doing. When you copy from Word your clipboard is carrying two versions of that selection: the words, and the markup behind them. A plain text box normally takes the words and quietly bins the markup, which would be unhelpful here, so this page asks for the markup by name instead.
If you paste actual HTML source there is no second version to ask for, and it simply arrives as typed. Either habit works. Nothing is uploaded either way: the cleaning happens in your browser, and the developer tools Network tab will confirm it.
The parts Word wrote for Word: mso- properties, class names like MsoNormal, Office namespace tags such as <o:p>, conditional comments, font tags, empty paragraphs and the inline styles that pin every word to Calibri 11pt. Your headings, paragraphs, lists, links, tables and emphasis stay.
Either. When you copy from Word or Google Docs your clipboard carries the markup as well as the words, and this page asks for the markup rather than letting the text box throw it away. If you paste HTML source instead, there is no second version to ask for and it arrives exactly as typed.
Because Word does not write lists. It writes ordinary paragraphs with a bullet character glued to the front and a note in the style attribute saying which list and which level it meant. A cleaner that strips styles without reading that note leaves you with paragraphs that begin with a middle dot. This one reads the note and rebuilds real <ul> and <ol> elements, nesting included. Untick the option if you would rather keep the paragraphs.
Yes, and there is a specific trap it avoids. Docs expresses bold and italic as inline styles and nothing else, so removing the styles first would silently delete every emphasis in the document. Styling is promoted to real <strong> and <em> tags before anything is thrown away. The invisible wrapper Docs puts around a copied selection is removed too.
Usually not. The three ticked boxes are the ones almost everyone wants, and the result you get without touching anything is the one that suits a CMS: headings, lists, links, tables, bold and italic kept, everything Word wrapped around them gone. The one unticked box is for a less common job: tick it when you want the outline of the document and nothing else, because you intend to do the formatting yourself. Bold, italic and underlining are all kept until you ask for them to go.
The front page. This tool is for markup you want to keep in better condition; the front page is for text you want the markup taken off altogether.
No. The parsing and cleaning happen in your browser, in JavaScript, and nothing is posted to this server or anywhere else. If you want to check, open your developer tools, switch to the Network tab and clean something. It matters here more than on most pages, because a document pasted into a cleaner is usually one that has not been published yet.