An extension to my previous post. I played a bit with the DT web clipper trying to capture content from sites that require login. I think I understand, what’s the DT problem. When I launched the Clip to DT action, instead of the page content (as I would expect) the login screen was captured!
IMO the DT clipper just sends the page URL to the DevonThink application and the real parsing/capturing of the page occurs later in the app. And the app must go through the login process again because from the server point of view it is a new session. Even launching Clip to DT in the internal DT browser does not work! Your HTML to Markdown converter starts from scratch ignoring the existing session.
If it is so, and I am pretty sure it is, than I must say it’s totally wrong solution. No wonder that it does not work. Until you parse the page content directly in the browser using JavaScript and send the converted result as the markdown document to the DT via an API, you have no chance to create a good clipper.