Hi @jordesign! I consider @manton the authority and will defer to him if I'm wrong about any of this, but I have some thoughts:

Video and audio media?

As long as the file is part of the media library on the site and not hotlinked, I think it can and should be included in the media folder along with the images.

Any kind of metadata or custom fields related to content

Yeah, I think these can be added to the JSON pretty easily nested under a custom-fields object with field-name and field-value properties. For the HTML we'll want to mark them up with something like HTML5 data attributes (data-*), microdata, or RDFa.

Does it distinguish between post types?

By default it looks like it is primarily just for Posts, but I don't see why it can't distinguish between post types. We could handle them like this:

• For each post type, have a separate top-level folder with the post type. Inside each folder include the JSON and HTML. For the JSON, add a post_type property.
• All of the media can still live in the top-level media folder and have a relative link.

That sounds right @cagrimmett. One of the original reasons to create Blog Archive was to have a standard place for media, so that posts, metadata, and media could all be packaged together.

For custom fields and even standard WordPress fields, it should be straightforward to have those as part of the HTML or even just in the JSON. WordPress importers could always prefer the JSON.

JSON Feed has a convention of using underscores for extensions, so the data for an item might look something like:

{
	"id": "http://example.org/2024/...",
	"content_html": "<p>...",
	"date_published": "2024-03-01T12:24:30-06:00",
	"url": "http://example.org/2024/...",
	"_wordpress": {
		"post_parent": 0,
		"post_type": "post",
		"comment_status": "closed"		
	}
}

Those are just a few fields I noticed while skimming through a WXR export. As an aside, it would be interesting to look into which fields are currently used by the WordPress importer, and whether anything is missing from WXR that could be added in a new format. Just curious how complete WXR is (minus the attached media files).