KamenReader
commented
3 years ago First commit used a sloppy workaround for the content type (it actually did a full get; forgot to change that after confirming it worked). The updated one just changes the original code to allow redirects and should be more efficient.
The correct URL (or at least a redirect to it) for full-size blog post images is actually under "original_url" rather than "url" in the json. Unfortunately, the prior version of process_content_type won't work on it since it doesn't actually link directly to the image (it won't detect the mimetype, and thus won't assign an extension). I modified process_content_type slightly so it should now work as intended in all cases.