For
Vyra for faceless channels: narration, b-roll, and captions without showing your face
Vyra builds faceless videos from your narration and your own b-roll or screen recordings, placing footage where the script mentions it, captioned.
By Sulan Zhang, co-founder of Vyra · Updated 2026-09-24
A faceless channel comes down to two things. How fast you can cover a script with the right footage, and whether the captions carry it when the sound is off. Vyra reads your narration, knows what's in every clip you uploaded, and cuts the two together from a written brief. You supply the b-roll or screen recordings. It does the placement.
What you make
- Faceless narrated videos
- Explainers
- Tutorials and app walkthroughs
- B-roll montages under music
Your three most-used prompts
Narration to video
Use the voiceover as the spine. Cut the b-roll over it so each clip matches what I'm describing. When I say "morning routine" use the kitchen and coffee clips, when I say "the app" use the screen recording. Change shot every 3-5 seconds. 4-word captions, white, bold, lower third. No music.
Screen recording explainer
Make a 60-second 9:16 explainer from the screen recording and the voiceover. Crop the screen to the part I'm pointing at each time. Title card for the first 2 seconds that says "3 settings you should change". Word-by-word captions, yellow highlight on the word I'm saying.
Daily batch
Same style as yesterday's video. Caption font, placement, cut rhythm, end card. Build today's from the new voiceover and the clips in the "day 12" folder.
A typical workflow
- Record narration, one file per video. Upload it with your b-roll or screen recording.
- Describe the structure. What plays under which part, caption style, music or none.
- Review the cut. Fix one thing per message. "Swap the clip at 0:14 for the drone shot."
- Lock the style and describe the next video as "same style."
- Export 9:16 for Shorts and TikTok, 16:9 for YouTube, from the same edit.
What Vyra does that matters for you
- The agent hears "the app" and cuts to the clip it identified as the app.
- Scene analysis on every clip, so "the drone shot over the harbor" finds the right file.
- Captions word-by-word or in phrases, with emphasis words. Faceless viewers often watch muted.
- Reference matching copies the rhythm and text style of a channel you like.
- Bring your own Claude or ChatGPT over MCP from $24/mo, or use the built-in AI.
What it does not do
- It doesn't generate b-roll, stock, AI voices, or avatars. You bring the visuals and the voice.
- It doesn't write your script.
Example
Example creator (TODO)
FAQ
Do I need to record narration first? It works best that way. The narration sets the timing. You can also start from footage and add voiceover later.
Can it use stock footage I already licensed? Yes. Upload it like any other clip.
Can I make a 10-minute YouTube video, not just Shorts? Yes. Describe the sections and it builds the long cut. Then ask for three Shorts from the strongest moments.