Can You Make a Full-Body AI Avatar Video?
Most AI avatar tools animate a portrait, so the frame stops at the shoulders. Here is how a head-to-toe character sheet lets you frame a talking scene full length, and where that still breaks.
HexUGC is an AI influencer generator: you fix one character once, then generate vertical talking-to-camera videos from it. The character sheet it builds is head to toe rather than a portrait, and the stills you generate from it are full-height 9:16 frames, so a scene can be framed full length, waist up or close in. Framing is a decision you make one step before the video, and that is the part most tools do not give you.
Why most avatar tools stop at the shoulders
A talking-head generator takes one portrait and animates the mouth. The input is a bust, so the output is a bust. That is fine for a corporate explainer, where a floating head over a slide is the expected shape. It is wrong for creator video, where a good half of the message is carried outside the face: what the person is wearing, where they are standing, what they do with their hands, how they turn to camera at the start.
Crop into a shoulders-up frame and a short-form video reads as a webinar clip. Viewers scrolling TikTok have a very fast instinct for that, and it is the same instinct that makes them keep scrolling.
The board is head to toe, so the body exists
An avatar starts as a character board: one wide sheet showing your person front, left, right and back, head to toe, generated from a written description plus up to three optional likeness photos. Head to toe matters for more than turnarounds. It puts build, posture, footwear and the whole outfit on record, so later images have something to be consistent with below the neck. Why characters drift and what still slips through is covered in how to keep the same AI character in every video.
Framing happens at the still, not at the video
Once a board exists, create-image turns build a library of stills of your character: different settings, angles and outfits, each one anchored on the board rather than on the previous still. Every still comes back in 9:16, which is the shape the video pipeline wants.
Framing is part of what you ask for. "Standing in a kitchen, full length, morning light" and "waist up at a desk" are both just directions on the turn, and the still that comes back is what the clip inherits. A scene is one still plus a line of direction, so by the time you generate video the framing is already settled. Stills are 5 credits each and you keep the ones that look right, which makes trying three framings of the same idea cheap compared with generating three videos.
Motion reference keeps the framing you chose
If you want a performance rather than a mouth, motion reference drives a scene from a supplied clip, including one you pulled off TikTok. It is submitted so that the character's orientation follows your image, not the reference video, so a full-length still stays full length instead of being recropped to the source. That workflow is written up in turning a TikTok into an AI avatar ad.
What this does not do
Wide framing costs you lip sync. At full length the mouth is a handful of pixels, so the sync is technically happening but is not legible, and a voiced scene reads better as a mid shot. The usual pattern is to open on a full-length still, then cut to a closer scene for the lines that matter.
It does not make big movement reliable either. Fast motion, walking and long unbroken takes are where the current generation of models is weakest, which is part of why clips are short and why multi-scene projects are stitched from several of them. More on that in why AI videos look fake.
Generating several variants in one action is on our roadmap rather than shipped, so it is one video at a time. There is no publishing or scheduling, so you download the MP4 and post it yourself. Your first character board is free, one per account with no card, subject to a daily cap we fund, but videos are not free.
What it costs
A 10 second voiced video is 28 credits. Each library still is 5 credits, and a board turn is 5 if you have already used your free one. 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.
Frame the whole person
If you have been fighting a tool that only ever gives you a head, the constraint is upstream of the video: it never had a body to draw from. Build the character head to toe first and the framing question answers itself.