Faceless videos vs on camera: how to pick a format

A faceless video is a video where no one appears on camera: a narrator, real or AI, speaks over illustrations, animation, footage or text. The alternative is a person talking to camera, often called a talking head video, and many good videos mix the two. Each format earns attention a different way, so the right choice depends on what the video has to do. A useful way to decide is to ask, for each line of the script, whether the viewer needs to trust it or understand it.

Faceless videos: when the idea is the hero

Faceless videos work when what you're explaining matters more than who's explaining it: how-tos, lists, explainers, stories about a thing rather than a person. They're also the practical choice for posting often, since nobody has to film each one, and for anyone who'd rather not be on camera.

The risk is that they feel anonymous. Trust has to come from the script instead of a face, so it needs real numbers, named examples and a specific point of view. "Save money on rent" is anonymous. "Three lines to check in a lease before you sign" has a point of view.

Example: "Three lines to check in a lease before you sign", with illustrated scenes of a lease, a highlighter landing on each clause, and a calendar for the notice period.

Faceless doesn't have to mean stock footage. A handmade animation style gives a faceless channel a recognisable look:

Claymation, risograph and felt puppets: faceless styles from the ThreeCut homepage examples.

On camera: when trust is the point

A person on camera works when the question in the viewer's head is "who is telling me this, and do I believe them?" Opinions, founder stories, replies to customers, and announcements that affect people (a price change, a shutdown, a mistake) all belong here.

The risk is that a talking head goes flat after the hook. Two things prevent it: a tight edit (cut the pauses, the ums and the false starts) and captions on every word, so the eye keeps moving even when the picture doesn't. How to edit a talking head video walks through both.

Example: "Why we're raising prices, and what you get for it", told by the founder, looking at the camera.

Both: when you need trust and explanation

The mixed format puts a face where trust matters and visuals where understanding matters. In most videos that means the face opens and closes, because the opening and the ask are where the viewer decides whether to believe you, and the middle is where they need to see how something works.

Example: a cook on camera says, "If chopping onions makes you cry, your knife is the problem." The middle switches to illustration: a blunt knife crushes more of the onion's cells, which releases more of the compound that stings your eyes; a sharp knife slices cleanly, and a chilled onion releases it more slowly. Back on camera: "Sharpen your knife tonight, and put the onion in the fridge half an hour before you cut it."

The animation shows something a camera can't (what happens inside the onion), and the face makes the advice feel like it comes from someone who cooks. There are a few ways to lay out the mix:

Three ways to mix: you above the animation, a cutaway to animation and back, and you in a bubble over it. From the ThreeCut homepage examples.

A quick decision table

The same idea, three ways

The idea: how to keep a houseplant alive while you're away for two weeks, using a jar of water and a cotton string as a wick.

• Faceless, 30 seconds. An illustrated plant, a jar and a string. The narrator: "Two weeks away, one jar of water, one cotton string." The scenes show the string pushed into the soil, the other end in the jar, and water creeping along it.
• On camera, 30 seconds. You, in front of your own plants: "I travel every month and these are four years old. Here's the whole trick." You set up the jar and string in one take.
• Both, 45 seconds. You on camera for the hook. An illustrated close-up in the middle shows water climbing the string into the soil, which is hard to film and easy to draw. Back to you: "Test it for three days before you leave, so you know the string is long enough."

All three work. The faceless one is fastest to make, the on-camera one builds a following for you, and the mixed one explains best.

Pick the shape last

Once you know the format, pick the shape for where the video will live: 9:16 for Reels, TikTok and Shorts, 1:1 for feeds, 16:9 for YouTube and websites. Write for the smallest screen it will appear on, with big text and one thing on screen at a time.

Common questions

What is a faceless video?

A faceless video is one where the creator never appears on camera. The story is carried by narration (your own voice or an AI voice) over illustrations, animation, stock or screen footage, or on-screen text. Faceless videos are common for explainers, lists, how-tos and channels that post often.

Do faceless videos perform as well as videos with a face?

It depends on the job. Faceless videos do well when the idea or the visuals are the draw, such as explainers and how-tos. Videos with a person tend to do better when trust or personality is the draw, such as opinions, stories and announcements. Many creators mix both.

How do you make a faceless video?

Write a short script, then pair each line with a visual: an illustration, an animation, a clip or a screen recording. Add narration, either your own recording or an AI voice, and captions on every word. Keep one idea per video and one thing moving on screen at a time.

What is a talking head video?

A talking head video is a video of a person speaking to the camera, usually framed from the chest up. It's the standard format for vlogs, founder updates, interviews and many UGC ads. Its strength is trust; its weakness is that it needs tight editing to hold attention.

---

ThreeCut makes all three formats from the same script: Faceless animation, Face only, and Face + animation, in any of those shapes. If you're unsure which to pick, write the script first. Once you can see which lines ask for trust and which ask for understanding, it's usually clear which format fits.