Add flake.nix with portaudio + CUDA, microphone support in transcribe.py

flake.nix

@@ -0,0 +1,21 @@
+{
+  description = "Cohere Transcribe dev environment with microphone support";
+
+  inputs = {
+    nixpkgs.url = "github:NixOS/nixpkgs/nixpkgs-unstable";
+  };
+
+  outputs = { nixpkgs, ... }: {
+    devShells.x86_64-linux.default = nixpkgs.legacyPackages.x86_64-linux.mkShell {
+      buildInputs = with nixpkgs.legacyPackages.x86_64-linux; [
+        portaudio
+        cudaPackages.cudatoolkit
+      ];
+
+      shellHook = ''
+        export LD_LIBRARY_PATH="${cudaPackages.cudatoolkit}/lib:$LD_LIBRARY_PATH"
+        echo "Dev shell ready - microphone input enabled"
+      '';
+    };
+  };
+}

transcribe.py

@@ -1,3 +1,6 @@
+import sys
+import numpy as np
+import sounddevice as sd
 from transformers import AutoProcessor, CohereAsrForConditionalGeneration
 from transformers.audio_utils import load_audio
 from huggingface_hub import hf_hub_download
@@ -18,14 +21,31 @@ def transcribe_audio(audio, language="en"):
     text = processor.decode(outputs, skip_special_tokens=True)
     return text
 
-# Use demo audio file from Hugging Face
+def record_audio(duration, samplerate=16000):
-print("Loading demo audio...")
+    print(f"Recording for {duration} seconds...")
-audio_file = hf_hub_download(
+    audio = sd.rec(int(duration * samplerate), samplerate=samplerate, channels=1, dtype='float32')
+    sd.wait()
+    return audio.flatten()
+
+# Parse arguments
+if len(sys.argv) > 1 and sys.argv[1] == "--mic":
+    duration = int(sys.argv[2]) if len(sys.argv) > 2 else 5
+    try:
+        mic_audio = record_audio(duration)
+        print("Transcribing...")
+        text = transcribe_audio(mic_audio)
+        print(f"\nTranscription:\n{text}\n")
+    except OSError as e:
+        print(f"Microphone error: {e}")
+        print("Hint: Run with nix-shell for PortAudio support")
+else:
+    print("Loading demo audio...")
+    audio_file = hf_hub_download(
         repo_id="CohereLabs/cohere-transcribe-03-2026",
         filename="demo/voxpopuli_test_en_demo.wav",
-)
+    )
-audio = load_audio(audio_file, sampling_rate=16000)
+    audio = load_audio(audio_file, sampling_rate=16000)
 
-print("Transcribing...")
+    print("Transcribing...")
-text = transcribe_audio(audio)
+    text = transcribe_audio(audio)
-print(f"\nTranscription:\n{text}\n")
+    print(f"\nTranscription:\n{text}\n")