Formats · Length 30-90s short, 8-15 min long · 9:16 or 16:9 · beginner
Faceless video: how to shoot, edit, and prompt it
A faceless video tells a story or explains something with voiceover, b-roll, screen recordings, and text, and you never appear on camera.
By Sulan Zhang, co-founder of Vyra · Updated 2026-09-24
- Platforms
- YouTube, TikTok, Reels, Shorts
- Length
- 30-90s short, 8-15 min long
- Aspect
- 9:16 or 16:9
- Difficulty
- beginner
- Gear
- a mic, stock or own b-roll, screen recordings
The script is the product. Every sentence should name something you can show.
At a glance
| Best for | Niche education, finance, history, tech, people who don't want to be on camera |
| Hook style | The boldest line over the strongest visual |
| Pacing | 3-5 seconds per clip, matched to the voiceover |
| Caption style | Full phrase captions, key terms bigger |
| Music | Continuous bed, ducked under the voice |
| Shot list | B-roll for every noun in the script. Screen recordings. Stills with a slow zoom. Text cards for numbers. |
| Common mistakes | Stock footage that doesn't match the words. Flat voiceover. Same zoom on every still. |
Structure
- 0-3s: the boldest line over the best visual
- 3-10s: what the video will explain
- 10-70s: the body, visual changes on every new idea
- 70-85s: the takeaway
- 85-90s: the next video
How to shoot it
- Script first.
- Record the voiceover in a quiet room, close to the mic, with some energy.
- B-roll from your own footage first, screen recordings second, stock last.
- High-res screenshots.
- List every number in the script for text cards.
How to edit it
- Cut the voiceover tight first.
- A visual on every noun. No visual outlasts its sentence.
- Slow zoom on stills, switching direction.
- Numbers as big text when spoken.
- Captions on everything.
- Music 10 dB under the voice.
Prompt it in Vyra
Build a 75-second faceless video from my voiceover and the clips and screenshots in the project. Put a matching visual on every thing I mention, slow zoom on the screenshots, and show any number I say as big text. Full phrase captions, music low under my voice, and start on the line "most creators quit at video twelve".
Replace the stock clips that don't match what I'm saying with the closest match from my own footage.
Match the caption font and the text card style of the attached reference.
Variations
- Screen only, software and web content.
- Documentary style, archival stills and footage.
- Listicle, numbered items, one clip each.
FAQ
Is faceless still worth doing? For niche education, yes. The voice and script are the product and it's easy to repeat.
Where does the b-roll come from? Your own footage and screen recordings first. They do better than stock.
Can Vyra match clips to the script? Yes. It transcribes the voiceover and analyzes every clip, so it can put the laptop clip on the sentence about "editing at 2am".