Gemini Live API: print the transcripts

Gemini API
,
1

I’m using this Gemini Live API tutorial now: cookbook/quickstarts/Get_started_LiveAPI.py at c063f0dbf13aa0da5f4b75931e174f9e02f16bce · google-gemini/cookbook · GitHub

Is there a way, maybe a flag to set, to also print the transcript of our speech too?
AWS Nova Sonic shows that by default. I need something similar here too.

3

Hey @Tina_Jasmine , you can pass the input_audio_transcription parameter in the config to retrieve the input audio transcription.

config = {
    "response_modalities": ["TEXT"],
    "input_audio_transcription": {},
}

Please refer to the documentation for more details.

Thanks.

5

But when I add it, voice doesnt work anymore. How can I keep both?

7

Hey, I just checked on my end and it worked fine, You just needed to update the receive_audio function.

async def receive_audio(self):
    "Background task to reads from the websocket and write pcm chunks to the output queue"
    while True:
        turn = self.session.receive()
        async for response in turn:
            if data := response.data:
                self.audio_in_queue.put_nowait(data)
                continue
            if text := response.text:
                print(text, end="")
            if response.server_content.output_transcription:
                print("Transcript:", response.server_content.output_transcription.text)
            if response.server_content.input_transcription:
                print('Transcript:', response.server_content.input_transcription.text)

        while not self.audio_in_queue.empty():
            self.audio_in_queue.get_nowait()