Is there any way to reuse or get the content of audio input so it's not sent with every request?

I’m sending the audio from a microcontroller so ram/space is very limited. While I can store the audio and send it along each time the request continues to grow and grow and take longer on such a small device. Is there way to reuse the same audio already sent to continue a conversation on the server so I don’t have to resend the base64 audio every time?