Reference-led generation
Upload one clear character image and add a short brief for appearance, clothing, accessories, and intended use.
Product features
Everything You Need to Create and Drive a 3D VTuber
Virmii combines AI 3D model generation, VRM avatar management, real-time browser face tracking, and OBS live output in one VTuber software workspace.
From input to output
One 3D VTuber workflow from reference image to live output.
Create the model, follow VRM preparation, test its tracking capability, and move into streaming without stitching together a separate tool for every stage.
3D model pipeline
Tripo powers the initial 3D model generation and automatic humanoid rigging before Virmii standardizes the result, prepares avatar metadata, and builds the VRM.
Browser face tracking
Use a supported desktop browser to drive head movement, gaze, blinks, mouth shapes, and available expressions.
OBS Browser Source
Create a private live-stage address with transparent rendering for an OBS Browser Source scene.
Portable downloads
Download the completed VRM and, when available, the source GLB through short-lived private links.
Controlled asset access
Reference images use a short-lived public handoff for 3D generation, then move to owner-protected storage. Generated assets use time-limited links.
Designed for honest output
Tracking capability follows the model, not the marketing copy.
A model with validated face topology can support richer blinking, mouth shapes, and expressions. Other models fall back to reliable head and body movement. Virmii labels that capability instead of presenting every result as identical.
Full tracking
Head rotation, gaze, blinking, mouth shapes, and supported expressions when the generated facial rig passes validation.
Head-only fallback
Stable head and body motion when the source model cannot support meaningful facial deformation.
See the full workflow before you begin.
Learn what to upload, what each generation stage does, and how the result reaches OBS.