diff --git a/edge-tts.py b/edge-tts.py
index f149974..ec2efa4 100755
--- a/edge-tts.py
+++ b/edge-tts.py
@@ -5,6 +5,7 @@ import uuid
 import argparse
 import urllib.request
 import websocket # pip install websocket-client
+from nltk.tokenize import wordpunct_tokenize
 from xml.sax.saxutils import quoteattr as escape
 try:
 	import thread
@@ -81,7 +82,7 @@ if __name__ == "__main__":
 		volumeString = args.volume
 		sentenceBoundaryEnabled = 'True' if args.enable_sentence_boundary else 'False'
 		wordBoundaryEnabled = 'True' if args.disable_word_boundary else 'False'
-		text = args.text.replace(chr(9), " ").replace(chr(13), " ").replace(chr(32), " ")
-		run_tts()
+		for text in wordpunct_tokenize(args.text.replace(chr(9), " ").replace(chr(13), " ").replace(chr(32), " ")):
+			run_tts()
 	elif args.list_voices is True:
 		list_voices()