Create stt.py

stt.py

@@ -0,0 +1,14 @@
+from faster_whisper import WhisperModel
+
+# Load model once
+stt_model = WhisperModel("base", compute_type="int8")
+
+def speech_to_text(audio_file):
+
+    segments, _ = stt_model.transcribe(audio_file)
+
+    text = ""
+    for segment in segments:
+        text += segment.text
+
+    return text.strip()