Can I use only audio input API (without output) to recognize speech in streaming mode

Since output speech generation is a little bit too expensive https://community.openai.com/t/how-to-decrease-cost-spend-in-openai-realtime-console-demo/1049199 , can I use just audio input API to recognize speech in streaming mode? 