Problem with Speech-To-Text

Hi there,
I am using the heygen streaming-avatar 2.0.10 library in VoiceChat mode when working with an interactive avatar in React.
I have a problem with voice recognition which is as follows.
For example if I say the following text: “I'm sure that speaking aloud helps you.”
As a result of recognition I get in : “Iam surethat speaking aloudhelps you.”
This problem is quite often repeated in a conversation.

What should I do to solve the problem ?