I might be misinterpreting the problem, but couldn't you have KM use the information passed to it from the HTML prompt, process that information in KM in whatever way necessary, use KM to write an HTML file, using a skeleton HTML file with KM variables inserted into the parts that vary, then have KM open that HTML file in the browser of your choice? I routinely have KM write HTML by inserting variables into skeleton HTML, then have KM open the HTML file so I can interact with it, having javascript in the page process my choices and sending the values back to KM for further processing. That further processing could include writing HTML (using the variables sent from the first HTML), then having KM open the new HTML page to use that new information for whatever purposes you like. There is more information here:
1 Like