diff --git a/vvd-extractor/src/main/resources/python/tag_mp3.py b/vvd-extractor/src/main/resources/python/tag_mp3.py
new file mode 100644
index 00000000..fe680ceb
--- /dev/null
+++ b/vvd-extractor/src/main/resources/python/tag_mp3.py
@@ -0,0 +1,80 @@
+def add_tag(input_file, thumbnail_file, label_dict, info_dict, audio_extractor_name, audio_tagger_name):
+  """
+  real add tags
+  """
+  from mutagen.mp3 import MP3, EasyMP3, TextFrame
+  from mutagen.id3 import ID3, APIC, TXXX, ID3NoHeaderError
+  import imghdr
+
+  # Use EasyMP3 for simple id3 tags
+  file = EasyMP3(input_file, ID3=ID3)
+  file.delete()
+  file["title"] = info_dict["defaultName"]
+  file["artist"] = info_dict["artistString"]
+  file["date"] = info_dict["publishDate"]
+  file["genre"] = "VOCALOID"
+  file["comments"] = "All rights belong to {}".format(info_dict["artistString"].split("feat.")[0].strip())
+  file.save(v2_version=3)
+
+  # Use MP3 for more complex id3 tags
+  tags = ID3(input_file)
+
+  with open(thumbnail_file, "rb") as f:
+    image_type = imghdr.what(f)
+    thumbnail_data = f.read()
+
+  # Add APIC frame for album art
+  tags.add(APIC(3, 'image/' + image_type, 3, 'Front cover', thumbnail_data))
+
+  # Add TXXX frames for custom strings
+  txxx_names = ["downloaded by", "pv url", "extracted by",
+                "tags edited by", "tags provided by", "made by"]
+
+  txxx_values = [label_dict["downloaderName"],
+                 list(filter(lambda x: x["id"] == label_dict["vocaDbPvId"], info_dict["pvs"]))[0]["url"],
+                 audio_extractor_name,
+                 audio_tagger_name,
+                 "VocaDB (https://vocadb.net/S/{})".format(info_dict["id"]),
+                 "CXwudi's vocadb-video-downloader-new (https://github.com/CXwudi/vocadb-video-downloader-new)"]
+
+  for name, value in zip(txxx_names, txxx_values):
+    tags.add(TXXX(encoding=3, desc=name, text=value))
+
+  if info_dict["albums"]:
+    album_strs = map(lambda album: album["name"], info_dict["albums"])
+    tags.add(TXXX(encoding=3, desc="included by", text="Albums [{}]".format(", ".join(album_strs))))
+
+  tags.save(input_file, v2_version=3)
+
+# ...the rest of your original code that calls this function can remain the same
+
+
+def main():
+  import json
+  import argparse
+  # Set up CLI Arguments
+  parser = argparse.ArgumentParser()
+
+  parser.add_argument("-i", "--input", required=True, help="input audio file")
+  parser.add_argument("-l", "--label", required=True, help="input label file")
+  parser.add_argument("-t", "--thumbnail", required=True, help="input thumbnail file")
+  parser.add_argument("-if", "--info", required=True, help="input info file")
+  parser.add_argument("-aen", "--audio-extractor-name", required=True, help="audio extractor name")
+  parser.add_argument("-atn", "--audio-tagger-name", required=True, help="audio tagger name")
+
+  # Grab the Arguments
+  args = parser.parse_args()
+  with open(args.label, "rb") as f:
+    label_dict = json.load(f)
+  with open(args.info, "rb") as f:
+    info_dict = json.load(f)
+
+  # commented out due to a japanese wired encoding problem
+  # e.g. 霊々音頭でまた来世 is 霊々音頭て　゙また来世　in niconico website title
+  # print("received args are " + str(vars(args)))
+
+  add_tag(args.input, args.thumbnail, label_dict, info_dict, args.audio_extractor_name, args.audio_tagger_name)
+
+
+if __name__ == "__main__":
+  main()