What is Virmii?

Virmii is an AI VTuber maker and 3D avatar studio. It turns a single character reference image into a textured, rigged 3D model, standardizes it into a VRM avatar, and lets you drive it with browser-based face tracking. You can preview and calibrate the avatar, then send it to OBS through a private live-stage URL for streaming.

How to use Virmii?

1. Upload a clear, well-lit full-body character reference image (PNG, JPG, or WebP) and name the avatar. 2. Generate the base 3D model — Virmii uses Tripo for AI 3D generation and automatic humanoid rigging. 3. Let the pipeline standardize and validate the rig, evaluate facial capability, and build the VRM. 4. Preview and calibrate the avatar in the browser with camera access, tuning head, gaze, blink, mouth, and expression response. 5. Copy the private live-stage URL into an OBS Browser Source. 6. Download the completed VRM (and source GLB when available) or keep using the hosted browser stage.

Core features of Virmii

Reference-led AI generation: create a 3D avatar from a single character image
Automatic humanoid rigging and VRM preparation with rig validation
Browser face tracking for head movement, gaze, blinks, mouth shapes, and expressions
Head-only fallback tracking for models without facial rig support
Private OBS live-stage URL with transparent rendering for Browser Source
VRM model library with preview and management of saved avatars
Portable downloads of VRM and source GLB via time-limited private links
Credit-based generation with free, monthly, and credit-pack options