diff --git a/cosyvoice/cli/cosyvoice.py b/cosyvoice/cli/cosyvoice.py
index a7bfab4..fc1ea90 100644
--- a/cosyvoice/cli/cosyvoice.py
+++ b/cosyvoice/cli/cosyvoice.py
@@ -102,9 +102,9 @@ class CosyVoice:
                 yield model_output
                 start_time = time.time()
 
-    def inference_cross_lingual(self, tts_text, prompt_speech_16k, stream=False, speed=1.0, text_frontend=True):
+    def inference_cross_lingual(self, tts_text, prompt_speech_16k, zero_shot_spk_id='', stream=False, speed=1.0, text_frontend=True):
         for i in tqdm(self.frontend.text_normalize(tts_text, split=True, text_frontend=text_frontend)):
-            model_input = self.frontend.frontend_cross_lingual(i, prompt_speech_16k, self.sample_rate)
+            model_input = self.frontend.frontend_cross_lingual(i, prompt_speech_16k, self.sample_rate, zero_shot_spk_id)
             start_time = time.time()
             logging.info('synthesis text {}'.format(i))
             for model_output in self.model.tts(**model_input, stream=stream, speed=speed):
diff --git a/cosyvoice/cli/frontend.py b/cosyvoice/cli/frontend.py
index 8770e31..99cdb18 100644
--- a/cosyvoice/cli/frontend.py
+++ b/cosyvoice/cli/frontend.py
@@ -178,8 +178,8 @@ class CosyVoiceFrontEnd:
         model_input['text_len'] = tts_text_token_len
         return model_input
 
-    def frontend_cross_lingual(self, tts_text, prompt_speech_16k, resample_rate):
-        model_input = self.frontend_zero_shot(tts_text, '', prompt_speech_16k, resample_rate)
+    def frontend_cross_lingual(self, tts_text, prompt_speech_16k, resample_rate, zero_shot_spk_id):
+        model_input = self.frontend_zero_shot(tts_text, '', prompt_speech_16k, resample_rate, zero_shot_spk_id)
         # in cross lingual mode, we remove prompt in llm
         del model_input['prompt_text']
         del model_input['prompt_text_len']