Every recording form has a switch called "Record my voice". "Record a talk-through" starts with it on. The other kinds start with it off.
Off: the AI narrates
You record in silence. When you save, Vistralo adds an AI voice: it finds the screens, writes a line for each and speaks it. This needs a cloud workspace. On the Free plan it also needs your own OpenAI and HeyGen keys.
With the switch off, the form can also take a "Website URL (optional)". If the recording shows a public website, give its address and Vistralo reads the page's code so the narration can name the real technology behind an effect.
On: you narrate
Your microphone is the audio in the recording. Saving it keeps your voice and does not start AI narration. Sound from the tab you are sharing is left out, so your voice is the only sound.
Which to choose
- Use your own voice when the reasoning is yours to give: why a design went this way, what you want the reviewer to look at.
- Use the AI voice when you would rather not speak, when the explanation should sound the same every time, or when recording until it sounds right would take longer than the task deserves.
Changing your mind afterwards
A recording saved without narration can still get it. Open the project and use "Add AI voice-over". If it already has one, the button reads "Redo AI voice-over".