diff --git a/octodns/idna.py b/octodns/idna.py
index bc91d46..bf30343 100644
--- a/octodns/idna.py
+++ b/octodns/idna.py
@@ -2,6 +2,8 @@
 #
 #
 
+from collections.abc import MutableMapping
+
 from idna import decode as _decode, encode as _encode
 
 # Providers will need to to make calls to these at the appropriate points,
@@ -12,6 +14,7 @@ from idna import decode as _decode, encode as _encode
 def idna_encode(name):
     # Based on https://github.com/psf/requests/pull/3695/files
     # #diff-0debbb2447ce5debf2872cb0e17b18babe3566e9d9900739e8581b355bd513f7R39
+    name = name.lower()
     try:
         name.encode('ascii')
         # No utf8 chars, just use as-is
@@ -34,3 +37,35 @@ def idna_decode(name):
         return _decode(name)
     # not idna, just return as-is
     return name
+
+
+class IdnaDict(MutableMapping):
+    '''A dict type that is insensitive to case and utf-8/idna encoded strings'''
+
+    def __init__(self, data=None):
+        self._data = dict()
+        if data is not None:
+            self.update(data)
+
+    def __setitem__(self, k, v):
+        self._data[idna_encode(k)] = v
+
+    def __getitem__(self, k):
+        return self._data[idna_encode(k)]
+
+    def __delitem__(self, k):
+        del self._data[idna_encode(k)]
+
+    def __iter__(self):
+        return iter(self._data)
+
+    def __len__(self):
+        return len(self._data)
+
+    def decoded_keys(self):
+        for key in self.keys():
+            yield idna_decode(key)
+
+    def decoded_items(self):
+        for key, value in self.items():
+            yield (idna_decode(key), value)
diff --git a/tests/test_octodns_idna.py b/tests/test_octodns_idna.py
index 0c6b125..2e09401 100644
--- a/tests/test_octodns_idna.py
+++ b/tests/test_octodns_idna.py
@@ -11,7 +11,7 @@ from __future__ import (
 
 from unittest import TestCase
 
-from octodns.idna import idna_decode, idna_encode
+from octodns.idna import IdnaDict, idna_decode, idna_encode
 
 
 class TestIdna(TestCase):
@@ -56,5 +56,87 @@ class TestIdna(TestCase):
         self.assertIdna('bleep_bloop.foo_bar.pl.', 'bleep_bloop.foo_bar.pl.')
 
     def test_case_insensitivity(self):
-        # Shouldn't be hit by octoDNS use cases, but checked anyway
         self.assertEqual('zajęzyk.pl.', idna_decode('XN--ZAJZYK-Y4A.PL.'))
+        self.assertEqual('xn--zajzyk-y4a.pl.', idna_encode('ZajęzyK.Pl.'))
+
+
+class TestIdnaDict(TestCase):
+    plain = 'testing.tests.'
+    almost = 'tésting.tests.'
+    utf8 = 'déjà.vu.'
+
+    normal = {plain: 42, almost: 43, utf8: 44}
+
+    def test_basics(self):
+        d = IdnaDict()
+
+        # plain ascii
+        d[self.plain] = 42
+        self.assertEqual(42, d[self.plain])
+
+        # almost the same, single utf-8 char
+        d[self.almost] = 43
+        # fetch as utf-8
+        self.assertEqual(43, d[self.almost])
+        # fetch as idna
+        self.assertEqual(43, d[idna_encode(self.almost)])
+        # plain is stil there, unchanged
+        self.assertEqual(42, d[self.plain])
+
+        # lots of utf8
+        d[self.utf8] = 44
+        self.assertEqual(44, d[self.utf8])
+        self.assertEqual(44, d[idna_encode(self.utf8)])
+
+        # setting with idna version replaces something set previously with utf8
+        d[idna_encode(self.almost)] = 45
+        self.assertEqual(45, d[self.almost])
+        self.assertEqual(45, d[idna_encode(self.almost)])
+
+        # contains
+        self.assertTrue(self.plain in d)
+        self.assertTrue(self.almost in d)
+        self.assertTrue(idna_encode(self.almost) in d)
+        self.assertTrue(self.utf8 in d)
+        self.assertTrue(idna_encode(self.utf8) in d)
+
+        # we can delete with either form
+        del d[self.almost]
+        self.assertFalse(self.almost in d)
+        self.assertFalse(idna_encode(self.almost) in d)
+        del d[idna_encode(self.utf8)]
+        self.assertFalse(self.utf8 in d)
+        self.assertFalse(idna_encode(self.utf8) in d)
+
+    def test_keys(self):
+        d = IdnaDict(self.normal)
+
+        # keys are idna versions by default
+        self.assertEqual(
+            (self.plain, idna_encode(self.almost), idna_encode(self.utf8)),
+            tuple(d.keys()),
+        )
+
+        # decoded keys gives the utf8 version
+        self.assertEqual(
+            (self.plain, self.almost, self.utf8), tuple(d.decoded_keys())
+        )
+
+    def test_items(self):
+        d = IdnaDict(self.normal)
+
+        # idna keys in items
+        self.assertEqual(
+            (
+                (self.plain, 42),
+                (idna_encode(self.almost), 43),
+                (idna_encode(self.utf8), 44),
+            ),
+            tuple(d.items()),
+        )
+
+        # utf8 keys in decoded_items
+        self.assertEqual(
+            ((self.plain, 42), (self.almost, 43), (self.utf8, 44)),
+            tuple(d.decoded_items()),
+        )