I used until today my own HTML parser to extract stuff (link, images, meta, etc) from a webpage. Not the fastest but it worked. However it is difficult to keep it up to date with the moving HTML standard. For example besides "src" now supports also "srcset".
I would like to switch to a 3rd party library but all the good ones seem to be based on Microsoft's IHTMLDocument. This might be a problem in the future if I decide to build for Mac/Linux.
There is any GOOD Delphi library that can parse HTML while being platform-independent? http://htmlp.sourceforge.net seems to be nice, but there is simply no documentation for it. And it is still beta.