Parser: Formula Type: Embeded

[bbappserver]

Very occasionally a website thinks it is a good idea to encode important data in two seperate formats that you need to suck out a file url.

For example the Tumblr API produces the json body

{
"...":"other keys we don't care abou",
"regular-body":"some encoded html containing an <img src=\"https://some.url/a.jpg\">"
}

Which is the actual CDN url for the data I'm interested in. I can awkwardly work around this by cutting out parts of the value until I'm left with the URL, but it would be nice if there was a recursive formula type, where the found output was passed to another parser and then all of the results collected.

[floogulinc]

Isn't this exactly what sub page parsers do?

[bbappserver]

If it is, it's either not working from JSON to html, or there is a lack of documentation on how to use it properly.