How to Make Short-Form Videos Without Showing Your Face
You can run a talking-to-camera account without ever being on camera. Here is how a reusable AI influencer does it, what stays anonymous, what does not, and what the whole thing costs.
HexUGC is an AI influencer generator. You design one character once, then generate vertical talking-to-camera videos of that character whenever you need them, and download each one as a 9:16 MP4. Nobody films anything, and the face on screen is not yours. That makes it one way to run a creator-style account while staying off camera yourself.
Faceless is two different jobs
Most advice about not showing your face means slideshow video: stock clips, screen recordings or still images cut to a voiceover. That works, it is cheap, and plenty of accounts are built on it. If that is the job you have, an assembler is the shorter route and you should use one, which is why we say the same thing in our comparison with timeline-style video tools.
The other kind of faceless is a person on camera who simply is not you. Short-form rewards a face talking directly at the viewer, and an account that never shows one is fighting the format. A character you generate keeps the format and removes you from it.
The character is the thing you build, not the clip
The mistake is treating every video as a fresh generation. Do that and you get a different person each time, and an account of strangers does not read as an account at all.
HexUGC starts with a board: one character sheet showing your character front, left, right and back, head to toe. That sheet is the anchor. Every still you create afterwards, in a kitchen, outdoors, in a different outfit, is generated against it, and every clip is generated against a still. The face carries across months of posting because nothing downstream is inventing it from scratch. What still slips through is covered in how to keep the same AI character in every video.
You describe the character in words. You can also upload likeness photos, but if the point is to stay unrecognisable, do not use photos of yourself. Describe someone who does not exist.
What making one video actually looks like
Pick a still from your library, say in a sentence what you want happening in the scene, and put your script in. Either write it yourself, paste one in whole, or give a freeform direction and let the model draft it for you to edit. ElevenLabs reads it in the voice stamped on the avatar, which you can change per video. Kling lip-syncs the still to that audio, ffmpeg stitches your scenes together, burns word-synced captions in, and hands back an MP4. Minutes, not a shoot.
If you want a specific performance, motion reference will drive the clip from a video you supply, including a TikTok URL. Your character does that movement, not the original person.
What stays anonymous, and what does not
The face and the voice are both synthetic, so neither identifies you. Your writing does not change, and your niche does not change, so anyone who knows your work will still recognise it. That is usually the point.
Two things are still yours. Disclosure is one: platforms have their own rules on labelling AI-generated content, and we cover the TikTok position in is AI-generated content allowed on TikTok. Comments are the other. A faceless account still needs someone answering as the character.
What this does not do
You generate one video at a time. There is no scheduling or publishing to TikTok, Instagram or YouTube, so you download the MP4 and upload it yourself. Long-form talking head is not the target either: this is short vertical video, and the cost scales with seconds of speech.
The videos do not get views on their own. Hook, niche and offer are still your job, and a synthetic face does not excuse a weak first three seconds.
What it costs
A 10 second voiced video is 28 credits, and each library still is 5. Your first character board is free, one per account, no card, subject to a daily cap; board turns after that are 5 credits. The first video is $5 for 50 credits, Starter at $15 a month is 300 credits and works out at roughly 10 videos, and Creator at $30 is 600 credits, about 21. Failed generations return their credits.