How to make FaceTuber videos: a YouTube presenter made by AI, no filming

Make presenter-led YouTube videos without a camera - a presenter animated by AI from one photo speaks your script, cut against footage of what they describe, edited for you.

The short answer

A FaceTuber video is a YouTube video with a person on camera who never filmed anything. In Stillscope you make the presenter once, from one clear head-and-shoulders photo. Then for each video you paste a script or upload narration, and Stillscope makes the finished video: the presenter speaks your words on camera with lips that match every word, and between appearances the video cuts to footage of what they are talking about. There is no timeline and nothing to edit.

Each video costs $0.65 a finished minute. A 20-minute video is $13.00. Making the presenter is a one-off extra, priced before you confirm.

Why a face helps a faceless channel

Advice, how-to and list videos sell trust, and viewers trust a person more than a voice over stock clips. A presenter gives the channel a face that comes back in every video, which is what makes a viewer feel they know the channel. Filming that person every week is what most faceless channels cannot do. A presenter made by AI removes the filming and keeps the face.

Step by step

  1. Make your presenter. Upload one clear head-and-shoulders photo: face the camera, good light, nothing covering the face. Use a photo of yourself or of someone who has agreed to it, never a celebrity or a stranger. You make the presenter once and use them in every video after that.
  2. Choose how they appear. Per channel, the presenter can sometimes share the screen beside a picture, or always be shown full screen.
  3. Write the script or upload the narration. Plain prose, written the way it should be spoken. Pick a voice from the library, use your own cloned voice, or upload narration you recorded. Long pauses in the narration are trimmed automatically.
  4. Pick a style if you want one. - Standard: the presenter keeps coming back through the video, with footage in between. - Relaxed: a full-pace opening, then mostly pictures, with the presenter back every two minutes. - Explainer: opens on pictures, shows every subject as documentary footage, and adds graphics for comparisons, numbers, maps and numbered parts.
  5. Check the price and press Create. You see the estimated length, the price and when the video will be ready. The median FaceTuber video took about 37 minutes to make once started.
  6. Download and upload. A finished 1080p video with its YouTube description written.

Where the footage comes from

Between the presenter's appearances, the video shows what they are describing: stock footage where it exists, and generated scenes where it does not. On the one project we measured closely, about three quarters of the cutaways were stock footage and a quarter were generated.

What it costs

Video lengthFaceTuber
10 minutes$6.50
15 minutes$9.75
20 minutes$13.00
30 minutes$19.50

That is the whole price per video: $0.65 a finished minute, charged on the finished video's real length, with no subscription and no monthly minimum. A video that fails costs nothing. The only other charge is making a presenter, once.

What FaceTuber channels make

The FaceTuber videos on our examples page run from 12 to 25 minutes and cover tips for shoppers, car care and Japanese home tips, each channel with its own presenter. Good subjects share one thing: a presenter can believably be the person giving the tips. Cars, home and garden, shopping, travel, history and numbered list videos all work.

One kind of subject does not: YouTube will not pay a channel whose AI-made presenter is presented as a human expert giving health, legal, financial or political advice. Keep your presenter a guide, not a doctor, lawyer or financial adviser. Can AI-made YouTube videos be monetized? has the details.

Follow YouTube's rules

The presenter is a realistic person made by AI, so when you upload, answer YouTube's question about altered or synthetic content honestly. And make every video its own: YouTube does not pay for mass-produced or repetitive content.

Questions

Do I need to film myself?

No. You upload one photo once, and the presenter speaks every script after that.

Can the presenter speak another language?

Yes. FaceTuber videos can be narrated in English, in twenty other languages written in the Latin alphabet, and in Arabic and Urdu. The Explainer style works in Latin-alphabet languages only.

Can I use my own voice?

Yes. Clone your voice once, or upload narration you recorded, and the presenter speaks with it.

How long can a FaceTuber video be?

As long as the script. Most run 10 to 25 minutes: the median FaceTuber video made with Stillscope ran 13.5 minutes.

More from the blog