diff --git a/nemo/collections/common/tokenizers/text_to_speech/tts_tokenizers.py b/nemo/collections/common/tokenizers/text_to_speech/tts_tokenizers.py
index bc6731ebcb88..5dc5c07de0c6 100644
--- a/nemo/collections/common/tokenizers/text_to_speech/tts_tokenizers.py
+++ b/nemo/collections/common/tokenizers/text_to_speech/tts_tokenizers.py
@@ -233,6 +233,77 @@ def __init__(
         )
 
 
+class GermanPhonemesTokenizer(BaseCharsTokenizer):
+    # fmt: off
+    PUNCT_LIST = (  # Derived from LJSpeech and "/" additionally
+        ',', '.', '!', '?', '-',
+        ':', ';', '/', '"', '(',
+        ')', '[', ']', '{', '}',
+    )
+    # fmt: on
+
+    def __init__(
+        self,
+        punct=True,
+        apostrophe=True,
+        add_blank_at=None,
+        pad_with_space=False,
+        non_default_punct_list=None,
+        text_preprocessing_func=german_text_preprocessing,
+    ):
+        """Deutsch phoneme-based tokenizer.
+        Args:
+            punct: Whether to reserve grapheme for basic punctuation or not.
+            apostrophe: Whether to use apostrophe or not.
+            add_blank_at: Add blank to labels in the specified order ("last") or after tokens (any non None),
+             if None then no blank in labels.
+            pad_with_space: Whether to pad text with spaces at the beginning and at the end or not.
+            non_default_punct_list: List of punctuation marks which will be used instead default.
+            text_preprocessing_func: Text preprocessing function for correct execution of the tokenizer.
+             Currently, it only applies lower() function.
+        """
+
+        de_ipa = "abdefhijklmnoprstuvwxyzçðøŋœɐɑɒɔəɛɜɡɪɹɾʃʊʌʒː̃"
+        de_suprasegmentals = "12"
+        super().__init__(
+            chars=de_ipa + de_suprasegmentals,
+            punct=punct,
+            apostrophe=apostrophe,
+            add_blank_at=add_blank_at,
+            pad_with_space=pad_with_space,
+            non_default_punct_list=non_default_punct_list,
+            text_preprocessing_func=text_preprocessing_func,
+        )
+
+    def encode(self, text):
+        """See base class."""
+        cs, space, tokens = [], self.tokens[self.space], set(self.tokens)
+
+        text = self.text_preprocessing_func(text)
+        for c in text:
+            # Add space if last one isn't one
+            if c == space and len(cs) > 0 and cs[-1] != space:
+                cs.append(c)
+            # Add next char
+            elif (c.isalnum() or c == "'" or c == "\u0303") and c in tokens:
+                cs.append(c)
+            # Add punct
+            elif (c in self.PUNCT_LIST) and self.punct:
+                cs.append(c)
+            # Warn about unknown char
+            elif c != space:
+                logging.warning(f"Text: [{text}] contains unknown char: [{c}]. Symbol will be skipped.")
+
+        # Remove trailing spaces
+        while cs[-1] == space:
+            cs.pop()
+
+        if self.pad_with_space:
+            cs = [space] + cs + [space]
+
+        return [self._token2id[p] for p in cs]
+
+
 class EnglishPhonemesTokenizer(BaseTokenizer):
     # fmt: off
     PUNCT_LIST = (  # Derived from LJSpeech and "/" additionally
diff --git a/nemo/collections/tts/torch/tts_tokenizers.py b/nemo/collections/tts/torch/tts_tokenizers.py
index 88a9d1df22f1..0cac295b6c06 100644
--- a/nemo/collections/tts/torch/tts_tokenizers.py
+++ b/nemo/collections/tts/torch/tts_tokenizers.py
@@ -20,5 +20,6 @@
     EnglishCharsTokenizer,
     EnglishPhonemesTokenizer,
     GermanCharsTokenizer,
+    GermanPhonemesTokenizer,
     IPATokenizer,
 )