diff --git a/embed.fnc b/embed.fnc
index cc2b9ac7aaaf..fbb3ed84f0ca 100644
--- a/embed.fnc
+++ b/embed.fnc
@@ -800,6 +800,9 @@ Adp	|U8 *	|bytes_to_utf8_free_me					\
 				|NN const U8 *s 			\
 				|NN STRLEN *lenp			\
 				|NULLOK void **free_me
+Adip	|U8 *	|bytes_to_utf8_temp_pv					\
+				|NN const U8 *s 			\
+				|NN STRLEN *lenp
 AOdp	|SSize_t|call_argv	|NN const char *sub_name		\
 				|I32 flags				\
 				|NN char **argv
diff --git a/embed.h b/embed.h
index 883b2fb72bdb..28d55b529cb5 100644
--- a/embed.h
+++ b/embed.h
@@ -157,6 +157,7 @@
 # define bytes_from_utf8(a,b,c)                 Perl_bytes_from_utf8(aTHX_ a,b,c)
 # define bytes_to_utf8(a,b)                     Perl_bytes_to_utf8(aTHX_ a,b)
 # define bytes_to_utf8_free_me(a,b,c)           Perl_bytes_to_utf8_free_me(aTHX_ a,b,c)
+# define bytes_to_utf8_temp_pv(a,b)             Perl_bytes_to_utf8_temp_pv(aTHX_ a,b)
 # define c9strict_utf8_to_uv                    Perl_c9strict_utf8_to_uv
 # define call_argv(a,b,c)                       Perl_call_argv(aTHX_ a,b,c)
 # define call_atexit(a,b)                       Perl_call_atexit(aTHX_ a,b)
diff --git a/inline.h b/inline.h
index bad3967673f4..0bac2329a888 100644
--- a/inline.h
+++ b/inline.h
@@ -1236,6 +1236,19 @@ Perl_bytes_to_utf8(pTHX_ const U8 *s, STRLEN *lenp)
     return bytes_to_utf8_free_me(s, lenp, NULL);
 }
 
+PERL_STATIC_INLINE U8 *
+Perl_bytes_to_utf8_temp_pv(pTHX_ const U8 *s, STRLEN *lenp)
+{
+    void * free_me = NULL;
+    U8 * converted = bytes_to_utf8_free_me(s, lenp, &free_me);
+
+    if (free_me) {
+        SAVEFREEPV(free_me);
+    }
+
+    return converted;
+}
+
 PERL_STATIC_INLINE bool
 Perl_utf8_to_bytes_new_pv(pTHX_ U8 const **s_ptr, STRLEN *lenp, void ** free_me)
 {
diff --git a/pod/perldelta.pod b/pod/perldelta.pod
index e0c3a7a72c16..35b6362898e3 100644
--- a/pod/perldelta.pod
+++ b/pod/perldelta.pod
@@ -360,7 +360,16 @@ well.
 
 =item *
 
-XXX
+Two new API functions are introduced to convert strings encoded in
+native bytes format to UTF-8.  These return the string unchanged if its
+UTF-8 representation is the same as the original.  Otherwise, new memory
+is allocated to contain the converted string.  This is in contrast to
+the existing L<perlapi/C<bytes_to_utf8>> which always allocates new
+memory.  The new functions are L<perlapi/C<bytes_to_utf8_free_me>> and
+L<perlapi/C<bytes_to_utf8_temp_pv>>.
+L<perlapi/C<bytes_to_utf8_temp_pv>> arranges for the new memory to
+automatically be freed.  With C<bytes_to_utf8_free_me>, you are
+responsible for freeing any newly allocated memory.
 
 =back
 
diff --git a/proto.h b/proto.h
index 513965b4ef88..3b125616efa0 100644
--- a/proto.h
+++ b/proto.h
@@ -9644,6 +9644,11 @@ Perl_bytes_to_utf8(pTHX_ const U8 *s, STRLEN *lenp);
 # define PERL_ARGS_ASSERT_BYTES_TO_UTF8         \
         assert(s); assert(lenp)
 
+PERL_STATIC_INLINE U8 *
+Perl_bytes_to_utf8_temp_pv(pTHX_ const U8 *s, STRLEN *lenp);
+# define PERL_ARGS_ASSERT_BYTES_TO_UTF8_TEMP_PV \
+        assert(s); assert(lenp)
+
 PERL_STATIC_INLINE void
 Perl_clear_defarray_simple(pTHX_ AV *av);
 # define PERL_ARGS_ASSERT_CLEAR_DEFARRAY_SIMPLE \
diff --git a/utf8.c b/utf8.c
index 928b6242a179..f2e8b80c7b4a 100644
--- a/utf8.c
+++ b/utf8.c
@@ -3256,6 +3256,7 @@ Perl_bytes_from_utf8(pTHX_ const U8 *s, STRLEN *lenp, bool *is_utf8p)
 /*
 =for apidoc      bytes_to_utf8
 =for apidoc_item bytes_to_utf8_free_me
+=for apidoc_item bytes_to_utf8_temp_pv
 
 These each convert a string C<s> of length C<*lenp> bytes from the native
 encoding into UTF-8 (UTF-EBCDIC on EBCDIC platforms), returning a pointer to
@@ -3275,6 +3276,13 @@ already there.
 In both cases, the caller is responsible for arranging for any new memory to
 get freed.
 
+C<bytes_to_utf8_temp_pv> simply returns a pointer to the input string if the
+string's UTF-8 representation is the same as its native representation, thus
+behaving like C<bytes_to_utf8_free_me> in this situation.  Otherwise, it
+behaves like C<bytes_to_utf8>, returning a pointer to new memory containing the
+conversion of the input.  The difference is that it also arranges for the new
+memory to automatically be freed by calling C<L</SAVEFREEPV>> on it.
+
 C<bytes_to_utf8_free_me> takes an extra parameter, C<free_me> to communicate.
 to the caller that memory was allocated or not.  If that parameter is NULL,
 C<bytes_to_utf8_free_me> acts identically to C<bytes_to_utf8>, always