If you’re using Gradio’s built-in browser mic input (e.g. gr.Audio(source="microphone") for real-time ASR), you can drop both sounddevice and pyaudio. You’ll still need ffmpeg on the server (or your dev box) so Gradio can transcode incoming audio blobs

requirements.txt

@@ -13,12 +13,7 @@ accelerate
 langchain
 langchain-community
 
-# audio input (choose at least one):
-sounddevice       # cross-platform realtime mic capture
-pyaudio           # alternative on Windows
-
-# (system requirement, not a pip package)
-#   • ffmpeg must be installed on your machine for any audio I/O
-#     • Debian/Ubuntu: sudo apt install ffmpeg
-#     • macOS (homebrew): brew install ffmpeg
-#     • Windows: download from https://ffmpeg.org/download.html
+# system requirement for audio I/O (ffmpeg must be installed):
+#   • Debian/Ubuntu: sudo apt install ffmpeg
+#   • macOS (homebrew): brew install ffmpeg
+#   • Windows: download from https://ffmpeg.org/download.html