Camera, Tripod, Teleprompter and the Uncut Take
The minimum technical setup for an AI video clone: camera resolution, a fixed tripod, a teleprompter, the uncut take, formats and why the audio can stay simple.

A clone is built from footage, so the footage has to be stable, continuous and unedited. These are the minimum technical rules.
Camera
A recent phone or a camera that records at 2K resolution at least. The guide I work from names the iPhone 14 Pro Max or later as the starting point.
Stabilisation
A tripod or a fixed support is mandatory.
Teleprompter
I strongly recommend a teleprompter app on a phone or tablet, so you can read the script while your eyes stay fixed on the lens.
Face and eyes
- Avoid turning your face, touching it, or looking anywhere but the camera.
- Speak directly to the camera, with natural movement.
Do not cut the takes
This is the rule that matters most. Even if there are mistakes, the recording has to be continuous. Do not edit and do not send fragmented clips. The raw material is required from start to finish. It is strictly prohibited to cut a take while you record: if you slip in the script, carry on.
Resolution and format
- Vertical (9:16): for ads and social media.
- Horizontal (16:9): for courses and long-form videos.
- Ideal: record the same shots in both formats, for maximum versatility.
Audio
The audio of these recordings is not used for the final avatar. It is only used for lip-sync. You do not need a professional microphone or a completely silent room, as long as the voice is minimally audible.
Checklist
- [ ] Phone or camera at 2K or better.
- [ ] Tripod or fixed support.
- [ ] Teleprompter app ready, script loaded.
- [ ] Takes recorded start to finish, with no cuts and no edits.
- [ ] Vertical and horizontal versions of the same shots.



