Avatar & voice studio
The studio for talking-head presenter video and voice, stock and custom avatars, digital-twin consent, a voice library, and voice cloning. Presenter work bills per minute from its own avatar pool.
The Avatar and voice studio is where you turn a written script into a talking-head presenter video, a chosen avatar speaking your words in a chosen voice. The studio brings together an avatar picker with stock and custom libraries, a digital-twin consent flow, and a voice library. Presenter work bills per minute from its own dedicated avatar pool.
What the studio makes
You write a script, pick an avatar and a voice, and generate a short presenter clip. Unlike image and video prompts, a presenter script is spoken verbatim. There's no prompt rewriting step, so the avatar says exactly what you typed. You can set a solid-color or image background to match your brand.
Avatars
There are four kinds of avatar:
- Stock avatars, a ready-made library you can use immediately, with no training.
- Photo avatars, created from a single uploaded photo.
- Digital-twin avatars, trained from a short recording of a real person, so the presenter looks like them.
- AI-generated avatars, described in words and generated from your prompt, with no real person involved.
Stock avatars are ready to speak right away; photo, digital-twin, and AI-generated avatars are custom-created, which runs in the background for a few minutes before the avatar is ready to use.
Voices
A voice library gives you a catalog of voices to pick from. Beyond that, voice cloning creates a custom voice from a short audio sample (around 30 seconds or more). Like custom avatars, a cloned voice trains asynchronously and becomes selectable once it's ready. A clone itself isn't billed to create. It only draws usage when it actually speaks in a generation.
Consent and digital twins
Because some custom avatars and cloned voices reproduce a real person's likeness, those are consent-gated, specifically photo and digital-twin avatars, and voice clones. (An AI-generated avatar described from a prompt depicts no real person, so it needs no consent.) Before you can create a photo or digital-twin avatar you capture a consent recording, and before you clone a voice you affirm you have the right to that person's likeness. The consenting person's details stay inside AnyBot's own records and are never sent to the generation provider, only the minimal signal needed to proceed crosses the boundary. Creating, and later deleting, a custom avatar or voice is audited.
How it runs and bills
Presenter generation rides the same fire-and-forget pipeline as every other media type: you request a clip, get an immediate acknowledgement, watch live progress, and end up with a finished file in your Media Drive. Your agent's Materials Library and Brand Kit can supply the branding around it.
Avatar and presenter work is metered per minute on its own dedicated AnyBot Tokens pool, separate from your video quota, see Media billing.
FAQ
Does the avatar say exactly what I write? Yes, the script is spoken verbatim, with no prompt rewriting.
Can I make an avatar of myself or a colleague? That's the digital-twin path, and it requires a captured consent recording before training. A cloned voice similarly requires a likeness affirmation.
How is it billed? Per minute, on a dedicated pool that never draws down your video quota.