So I got the ai llm api, and the stt api and and when i click system audio capture it doesn't seem like it reads the screen it just listen to the audio am I using it wrong?