How to Make AI Influencer Photos You Can Download
An AI influencer is a set of images before it is a video. Here is how the character board, the library turns and the downloads work on HexUGC, what each step costs, and where it stops.
HexUGC is an AI influencer generator: you design one character once, then produce images and vertical talking-to-camera videos of that character. The images come first, and plenty of people want them on their own. Your first character board is free, one per account with no card, each library still after that is 5 credits, and every image you keep can be downloaded as a file.
The hard part is not one photo, it is the tenth
One good AI portrait stopped being difficult a while ago. Type a description into almost any image model and you get a convincing face. The job an AI influencer actually needs is a set: the same person in a kitchen, outdoors, at a desk, in three different outfits, across months of posting.
Most image tools restart from the prompt on every generation, so the face is re-invented each time. Hair length moves, the jawline softens, the age slides by five years. Individually each picture is fine. As a grid it reads as several people, and an account of strangers does not read as an account at all. Why drift happens and what still slips through is covered in how to keep the same AI character in every video.
The board is the anchor, and the first one is free
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. That sheet is the reference everything downstream is generated against.
The first board is free, one per account, no credits and no card, bounded by a daily cap we fund, so on a busy day the free option can be absent and back the next morning. Regenerate as much as the attempt cap allows. Each attempt starts fresh from your description rather than from the previous sheet, because running AI over AI output compounds artefacts. Older sheets stay selectable, and if you already have a character sheet made elsewhere you can upload that instead.
Each photo is a turn against that board
Once a board exists, create-image turns build your library. You write what you want, the board is always sent as the base image, and you can attach earlier library images as numbered references when a specific outfit or setting needs to carry over. Every still comes back in 9:16 and costs 5 credits.
A turn is a candidate, not a commitment. You keep the ones that work and they join the library permanently; the rest are discarded. Framing is part of what you ask for, so full length, waist up and close in are just words in the direction, which is discussed further in making a full-body AI avatar video.
Getting the files out
The library page lists every kept image with a download button on each one, and they save straight to disk rather than opening in a tab. You can also upload your own real photos into the same library, which is useful when a genuine product shot or location photo belongs in the set.
That same library is what the video step reads from. A scene is one still plus a line of direction, so the photos you made for a grid are the photos your videos are built on, with nothing to redo. Turning one of them into a talking clip is covered in making your AI character talk.
What this does not do
Generated stills are 9:16 only. There is no square or 4:5 output, so a grid post means cropping the file yourself, and the board itself is a wide 16:9 sheet rather than a portrait.
It is not a photo editor. No inpainting, no retouching, no masking part of an image and changing only that. If something is wrong you re-run the turn with a clearer description. It is also not a face swap onto a photo you already have.
Generating several variants in one action is on our roadmap rather than shipped, so it is one turn at a time. There is no publishing or scheduling to Instagram or TikTok either. You download and post them yourself.
What it costs
A board turn is 5 credits once you have used your free one, and a library still is 5. A 10 second voiced video is 28 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 automatically.
Build the person, then the pictures
If your images keep coming out as different people, the problem is upstream of any single prompt: nothing is anchoring them. Fix the character once and the set follows.