transcription problem

I’ve encountered an issue while using Blabby.ai for dictation.

When I speak into the microphone (whether to ask a question or dictate a message), the app doesn't output my exact spoken words into the text field. Instead, it processes my speech through an AI model and inserts the generated AI answer/response directly into the input field.

Expected behavior:

The tool should transcribe my voice directly (pure Speech-to-Text) and insert what I actually said into the active text field.

Current behavior:

The tool takes my voice input, interprets it as a prompt, and generates an AI response instead of transcribing my words.

Could you please check if this is a bug with the current mode processing or if there’s a setting to force raw Speech-to-Text transcription without LLM generation?

Thanks for your help!