One thought that occurred to me several times in the last year or so is that roughly the level of abstraction offered by NestedText might make sense as part of a hierarchy of abstractions that could be built on.
We already have that with text files. Because end-of-line character combinations are special, text files in a given encoding are already more structured than streams of characters.
So, assuming UTF-8 character encoding:
CharacterStream
Text (EOL character combinations)
NestedText
MyNestedTextFormat (domain specific semantics)
With non-text files, this has already happened more than once. For example, both Zip files and Sqlite files are used as base formats for specifying other formats.