app.py

import gradio as gr
from vosk import Model, KaldiRecognizer
import wave
import json

# Set up the Vosk model
model = Model("vosk-model-small-uz-0.22")

def recognize_from_file(audio_file):
    wf = wave.open(audio_file, "rb")
    if wf.getnchannels() != 1 or wf.getsampwidth() != 2 or wf.getcomptype() != "NONE":
        return "Audio file must be WAV format mono PCM."

    recognizer = KaldiRecognizer(model, wf.getframerate())
    result_text = ""

    while True:
        data = wf.readframes(4000)
        if len(data) == 0:
            break
        if recognizer.AcceptWaveform(data):
            result = json.loads(recognizer.Result())
            result_text += result.get('text', '') + " "

    final_result = json.loads(recognizer.FinalResult())
    result_text += final_result.get('text', '')

    return result_text


iface = gr.Interface(
    fn=recognize_from_file,
    inputs=gr.Audio(type="filepath"),
    outputs="text",
    title="Speech Recognition from Audio File",
    description="Upload a WAV file for recognition."
)

iface.launch()