Think about the last product video that actually convinced you. There is a decent chance it had a moment where the person stopped talking to the camera, turned it round, and showed you the thing. That turn is where the video stops being a claim and starts being evidence.

Now try to film it with a script. You are reading from the front camera. You reach the thing you want to show. And to show it you have to stop recording, flip, and start again — which means two files, a join, and an edit you did not want to do.

Why switching normally means stopping

When an ordinary camera app changes camera, it tears down the capture session and builds a new one. The encoder stops with it, the file is finalised, and anything recorded afterwards is a second file. That is not a bug — it is simply how the pipeline is usually built.

Talk2Camera keeps the encoder running and changes only where the frames come from. The recording never ends, so there is no second file and no join. On hardware that cannot open both cameras at once, a held frame covers the moment the incoming camera takes to start, so the timeline never stalls. There is more detail on the camera switching page.

What this changes about how you write

Once switching is free, you write differently. You stop writing a monologue and start writing a route: talk here, walk there, show that, come back. The script becomes a plan for a journey rather than a wall of text, and the video gets better because you are pointing at real things instead of describing them.

Keep the walking sections short. Two or three sentences between stops is plenty — long enough to move, short enough that you are not reading while watching your feet.

And say the switch out loud in your script. "Let me show you" is a better transition than a jump cut, and it gives you a natural half-second to move the phone.

The teleprompter keeps up while you move, and everything you shoot arrives in the editor as one continuous take.