API/MCP creation of Weblink and Image objects

I’m building a Capacities workflow where ChatGPT helps process incoming research and reference material into structured objects.

At the moment, the API/MCP can create and update custom object types, but it cannot create native Weblink or Image objects. This creates two practical gaps:

A URL can be stored as text, but ChatGPT cannot create a native Weblink object, so Capacities’ preview/thumbnail generation is unavailable.

ChatGPT can relate an existing Image object to another object, but cannot upload/create the Image object itself.

It would be extremely useful if the API/MCP could:

create a native Weblink from a URL, with the same preview/metadata generation as the Capacities UI;

create/upload an Image object from supplied image bytes or a file reference;

then allow those native objects to be attached to custom object properties.

This would make external AI-assisted capture workflows much more complete: for example, browser/mobile capture → AI enrichment → structured Clipping object → project/area/tags → native preview image.

The existing object-relationship model already supports this very well; the missing piece is simply external creation of those native media objects.