Questions, answered plainly

No. The file is written straight into the folder you picked, and if a name is already taken it becomes Article-2.md rather than overwriting anything. There is no protocol handler and no browser dialog in between, so there is no step where a stray click drops the clip. The settings page also counts how many times you pressed the button and how many of those produced a note, so a silent gap would be visible.

No. Clean Clipper writes the .md file directly into a folder you picked once, using the browser file access API. Obsidian can be closed, uninstalled, or absent entirely – the file appears in the folder regardless, and Obsidian picks it up next time it opens the vault.

Because many clippers hand the note over through the clipboard, and on Linux under Wayland only the focused window may read the clipboard. If the target app is not focused at that moment, the body never arrives. Clean Clipper never puts the note on the clipboard to save it, so that failure mode does not exist here.

Yes, the same as anywhere else. The window manager, the clipboard and the desktop environment play no part in saving, because the file is written by the browser itself.

Yes, and it was measured. On a fifteen-table corpus twelve survived intact against seven for the engines compared. Tables are serialised by our own code rather than a generic converter, so a cell holding a list or a snippet of code no longer collapses the row. Merged cells are flattened, because Markdown has no way to express a merge.

No. Images are kept as ordinary Markdown links with no stray escaping, so they render as soon as you paste. What the extension does not do is download image files next to the note – the link points at the original.

It says so. On storefronts, feeds and search results there is no article to extract, and instead of handing you three hundred link labels the extension writes an extraction: no-article marker and stops. The decision is made by a measured threshold – when more than a quarter of the text sits inside link labels, it is navigation, not prose.

Your pages, no. Extraction and conversion happen entirely inside the browser, and no page text or title ever leaves it. Two things do: the address of a page where no article could be found, so the site can be taught, and the bare domain of a successful clip – never the path. Both stop the moment you switch them off in settings. Access to a page itself is granted by your own click and only for that tab.

Counters that live in your browser. They record how many clips you made, how much time that saved, which destination you used, and where an article could not be found. The report itself is shown to you in full and leaves only if you copy it and email it yourself; the panel also shows how many addresses have been sent, so what goes out is never out of sight.

No. A successful clip leaves behind its domain and nothing more – no path, no query, no title, so the list cannot show what you read on a site, only that you clipped there. Full addresses are kept in one case only: pages where extraction failed, because that is the one thing a fix cannot be written without. Everything kept is visible in the panel and can be cleared with one button.

When the page declares one, yes. It is read from JSON-LD, article:published_time, citation_date (the arXiv, PubMed and IEEE standard), time[datetime] and Unix stamps, in that order. When the page carries no date at all, the field stays empty rather than being filled with today. Some kinds of site – shops and timetables in particular – simply never publish a date.

Yes. Select something on the page and that is what gets clipped; select nothing and you get the whole page. The switch at the top of the window always shows which one you are looking at, and the behaviour can be turned off in settings.

No. Clean Clipper works on one page at a time and has no crawler and no batch mode. If you need dozens of pages in one go, this is the wrong tool.

No. There is no highlighter and no annotation layer. The extension converts what is already on the page and does not add anything to it.

No. The text is converted, not edited: no summarising, no reordering, no rewriting. What the page said is what the note says.

Chrome and other Chromium browsers: Edge, Brave, Vivaldi, Opera, Arc. It is built on Manifest V3, so it will not stop working when the old extension format is switched off.

Yes, entirely. Every feature described here is in the installed extension: no page limit, no account, no registration, no paid tier.

Write, with the address. The settings page has a report that already lists the sites where extraction failed, so copying it into the email is enough – no need to describe anything.