diff --git a/server/gpu/modal/reflector_diarizer.py b/server/gpu/modal/reflector_diarizer.py
index b1989a11..1a0e9f0a 100644
--- a/server/gpu/modal/reflector_diarizer.py
+++ b/server/gpu/modal/reflector_diarizer.py
@@ -11,7 +11,7 @@ from pydantic import BaseModel
 
 PYANNOTE_MODEL_NAME: str = "pyannote/speaker-diarization-3.0"
 MODEL_DIR = "/root/diarization_models"
-
+HUGGINGFACE_TOKEN = modal.Secret.from_name("my-huggingface-secret")
 stub = Stub(name="reflector-diarizer")
 
 
@@ -34,7 +34,7 @@ def download_pyannote_audio():
     Pipeline.from_pretrained(
         "pyannote/speaker-diarization-3.0",
         cache_dir=MODEL_DIR,
-        use_auth_token="***REMOVED***"
+        use_auth_token=HUGGINGFACE_TOKEN
     )