Advanced Voice-to-Text Transcription using Python and Whisper

import os, queue, json, sounddevice as sd, soundfile as sf, numpy as np, whisper
SAMPLE_RATE=16000
CHANNELS=1
MODEL_NAME="small"
DEVICE="cpu"

class Recorder:
    def __init__(self,sample_rate=SAMPLE_RATE,channels=CHANNELS):
        self.sample_rate=sample_rate;self.channels=channels;self.q=queue.Queue()
    def _callback(self,indata,frames,time,status):
        if status:print(status);self.q.put(indata.copy())
    def record(self,duration,out_path):
        with sd.InputStream(samplerate=self.sample_rate,channels=self.channels,dtype="int16",callback=self._callback):
            frames=[];import time;start=time.time()
            while time.time()-start<duration:frames.append(self.q.get())
            audio=np.concatenate(frames,axis=0);sf.write(out_path,audio,self.sample_rate)
            print(f"Recorded {duration}s to {out_path}")

class Transcriber:
    def __init__(self,model_name=MODEL_NAME,device=DEVICE):
        print("Loading model...");self.model=whisper.load_model(model_name,device=device)
    def transcribe(self,path):
        result=self.model.transcribe(path);return result["text"]

def save_text(text,path):
    os.makedirs(os.path.dirname(path),exist_ok=True)
    with open(path,"w",encoding="utf-8")as f:f.write(text)
    print(f"Saved transcript to {path}")

def main():
    print("1. Record new audio\n2. Transcribe existing file")
    choice=input("Choose (1/2): ")
    if choice=="1":
        dur=float(input("Enter duration in seconds: "))
        rec_path="recorded.wav";Recorder().record(dur,rec_path);file_path=rec_path
    else:
        file_path=input("Enter audio file path: ")
    transcriber=Transcriber()
    text=transcriber.transcribe(file_path)
    print("\n--- TRANSCRIPT ---\n",text)
    save_text(text,"transcript.txt")

if __name__=="__main__":main()