I expect several phases here:

(1) add a .dedent() method to str (and bytes?) - behaviors to consider mirroring are textwrap.dedent() and inspect.cleandoc().  Given their utility and similarities, it makes sense to offer both behaviors; behavior could be selected by a kwarg passed to the method.

https://docs.python.org/3/library/textwrap.html#textwrap.dedent

https://docs.python.org/3/library/inspect.html#inspect.cleandoc

(2a) Ponder the d" prefix - but in general I expect sentiment to be against yet another letter prefix.  They aren't pretty.  This would need a PEP.  Someone would need to champion it.

(2b) Ponder making cleandoc dedenting automatic for docstrings.  This would be enabled on a per-file basis via a `from __future__ import docstring_dedent` or similar as Serhiy suggested.  No prefix letter required.  Several releases later we could consider making it the default.  This will also need needs a PEP.

(3) Optimizations when .dedent() is called on a constant?  Nice to have, But I suggest we land (1) first as its own base implementation PR. Then consider the follow-ons in parallel.

I believe the current patch contains (1)+(3) right now.  If so we should simplify it to (1) and make (3) am immediate followup as saving the runtime cost and data space is worthwhile.

Ultimate end state: probably 1+2b+3, but even 1+3 or 1+2b is a nice win.