Speech-to-text - How does it work?

I am trying to use the speech to text API to write a pytest but I am having some issues. I am not clear if this is due to the configuration or my misunderstanding. Am I going about this correctly?

I am downloading wav files from here:

Then re-encoding it using this:

                audio_segment = AudioSegment.from_file(BytesIO(data))
                audio_segment = audio_segment.set_frame_rate(16000)
                audio_segment = audio_segment.set_channels(1)
                audio_segment = audio_segment.set_sample_width(2)
                wav_data = audio_segment.export(format="wav").read()
# Initialize RhasspyClient
client = RhasspyClient("http://localhost",12101,"/api")

This makes a call to the transcribe_wave file function I have

        result = await client.transcribe_wav_file(file)
        response_dict = json.loads(result)
        response = response_dict["text"]
        print(f"Transcribed text: {response}")

    async def transcribe_wav_file(self, wav_data: bytes) -> str:
        url = f"{self.url}/speech-to-text"
        headers = {"Content-Type": "audio/wav"}
        async with aiohttp.ClientSession() as session:
            async with session.post(url, headers=headers, data=wav_data) as response:
                text = await response.text()
            return text

The response should return “zero” but I am getting “is the garage door open” back as a result.

I made these changes to the profile.js, but not sure if they are correct.

cat ~/.config/rhasspy/profiles/en/profile.json 
    "dialogue": {
        "system": "rhasspy"
    "microphone": {
        "pyaudio": {
            "device": "10"
        "system": "pyaudio"
    "sounds": {
        "system": "aplay"
    "speech_to_text": {
        "kaldi": {
            "speech_to_text.kaldi.open_transcription": true
        "system": "kaldi"
    "text_to_speech": {
        "system": "nanotts"
    "wake": {
        "system": "porcupine"