Faceless video maker

What holds the screen instead of a person

Mostly type, moved deliberately. The statement role alone covers a single shouted line, three stacked words, a two line setup and payoff, a phrase cycling through several, an echoed ghost word, a broadcast lower third and a dictionary style definition. Around it sit counting numbers, building charts, ticking checklists, drawn funnels and cycles, a typing prompt bar and a conversation in message bubbles. None of it needs a camera and none of it needs stock footage.

The voice is doing more work than usual

With no face, the narration is the presence. Voices come from Google's Chirp3-HD set and are chosen by style and gender, in four styles: energetic, instructional, warm and neutral. The director is told to pick from the mood of your brief rather than the video type, so a sensitive subject gets warm or neutral even in a promo. In the builder you can go further and search the whole live voice catalogue per scene, plus set the delay, the playback speed and whether that scene captions itself.

Captions are not optional here

A faceless video is usually watched muted first. Captions are on by default and generated live from the narration, letters typing on slightly ahead of the voice in a fixed dark pill so they stay legible over any scene. They sit 44 pixels from the bottom in the wide cut and 120 in the vertical one, and cap at 72 percent of the frame width so a long line wraps rather than running edge to edge.

Where faceless is the wrong call

Two cases. Anything where trust comes from a specific person, a founder's apology, a surgeon explaining a procedure, a testimonial that needs a name and a face, is weaker without them. And anything where the product is a physical experience, food arriving at a table, a class in progress, reads flat as type and drawn graphics. Those want real footage or, in the second case, real photographs carrying the scenes.

How it works, in three steps

1. Step 1

   Decide the through line

   With no presenter, the script is the spine. Write the one sentence the viewer should repeat afterwards, and put it in the brief as the closing line.

2. Step 2

   Let it choose the voice from the mood

   Leave the voice unspecified in the brief and describe the tone instead. The style and gender are picked from that rather than defaulted by video type.

3. Step 3

   Check the muted pass

   Play it without sound in the preview. If a scene stops making sense, its on screen copy is leaning on the narration and needs rewriting.

The full walkthrough with screenshots is in the guide Add voiceover, captions and music to a video.

Limits worth knowing

• No footage is licensed in. Everything on screen is drawn by the engine, uploaded by you, or generated to a prompt.
• The synthesised voice is expressive but not directable line by line. There is no emphasis markup or pause control beyond delay and speed.
• Two languages do not have the premium voice tier at all and fall back to an older one: Malay and Filipino.
• Captions render as one centred pill. There is no per word karaoke styling, alternative placement or custom caption font.

Questions people ask

Related pages