VTuber is shorthand for Virtual YouTuber. In practice it means any creator who performs as a digital avatar animated in real time, whether they upload to YouTube, go live on Twitch, or do both. This wiki explains the word and the vocabulary around it so you can read chats, artist briefs, and debut posts without getting lost.
For a full explainer, read what a VTuber is. For a practical path into the hobby, use how to become a VTuber.
VTuber: definition and origin
A VTuber is an online personality who uses an animated avatar as their public face. The avatar is usually anime-inspired, but it can also be a mascot, a semi-realistic figure, or a simple PNG. Motion-capture or face-tracking software copies the performer’s expression onto the model while they talk.
The label comes from Japan in the mid-2010s. Virtual idols already existed; what made “Virtual YouTuber” stick was live, chat-facing performance. After early breakout talents, agencies industrialized the format and the word spread into English, Spanish, and other scenes.
Today the “YouTuber” part is historical. Plenty of VTubers are Twitch-first. The avatar-plus-live-performance idea is the definition that still holds.
Live2D
Live2D is the 2D animation system most associated with VTubers. An illustrator draws the character in layers (hair, brows, eyes, mouth, body, clothes). A rigger then wires those layers to tracking inputs so the character can tilt, blink, and lip-sync without becoming a 3D mesh.
People say “Live2D model” the way they say “PNG file”: it is both a file format family (often .json plus textures inside a folder or a .vtube.json setup) and a whole profession. Commission prices usually split art and rig.
3D VTubers
A 3D VTuber uses a three-dimensional character, closer to a game avatar. With a VR headset, full-body trackers, or a good IK setup, the talent can walk, dance, and share a 3D space with other models.
3D is heavier on the computer and usually more expensive to commission, but it is the standard for many concerts, 3D collabs, and “room” streams. Some talents keep a Live2D model for chatting and a 3D model for specials.
Tracking and motion capture
Tracking is how the computer knows what your face and body are doing. A webcam can follow the head and mouth. An iPhone can use ARKit for more stable face data. Full-body motion capture uses VR or dedicated suits/trackers.
Motion capture (mocap) is the broader term. In VTuber chat, “my tracking is cooked” usually means the face capture is lagging, jittering, or lost the mouth shapes.
Rigging
Rigging is the invisible skeleton of a model. The rigger decides how far the head turns, how the hair sways, which mouth shapes exist (visemes), and which hotkeys trigger jumps, tears, or a “dead” expression.
A beautiful illustration with a poor rig feels stiff. A simpler drawing with a good rig often looks more “alive” on stream. When you commission, you are buying both.
PNGTuber
A PNGTuber (or PNGtuber) uses one or more still images instead of a full Live2D/3D rig. Software can swap to an open-mouth image when the mic hears speech, or the creator hits a key.
PNGTubers are VTubers. The format is popular for tests, budget debuts, and joke alts. Many people debut as a PNGTuber while waiting on a Live2D queue.
VTuber agencies
An agency is a company that contracts VTuber talents, similar to a talent agency plus a production studio. The two names you will see most often:
- Hololive — Cover Corporation. Generations of talents, large 3D lives, and a very visible English branch (Hololive EN).
- Nijisanji — AnyColor. A wide roster, including English-speaking livers, with a strong variety-stream culture.
Other agencies and collectives exist in Japan, China, Korea, and the West. They differ on IP ownership, clip rules, and how much of the model the talent keeps if they leave. Always read the contract; the Twitter drama is not the documentation.
Indie means not signed. Indies still collab with agency talents sometimes; the word describes the business relationship, not the quality of the model.
Community words you will see in chat
- Oshi: the talent you support most. Oshi mark: the emoji or kaomoji that represents them in chat.
- Debut: the first official stream or video as that character. Redebut / 2.0: a new model or a return after a rebrand.
- Lore: the character’s backstory, canon, and running fiction.
- Clipper: someone who cuts stream highlights. Clip culture is how many VTubers get discovered.
- Anti: a hostile account. Not the same as criticism.
- Graduate / graduation: an agency talent’s farewell; they leave the company, and the character often retires with them.
- Liver: used especially in Nijisanji spaces as another word for the VTuber talent.
- Rig / model dump / VTS: shop talk for files and VTube Studio.
These words show up in memberships, Discord, and fan art tags. You do not need them to enjoy a stream. You need them to understand why chat is spamming a particular emoji.
Tools, in one list
- VTube Studio (VTS): popular Live2D tracking and hotkey app
- OBS Studio: mixes avatar, game, and alerts for the broadcast
- PRISM Live Studio: all-in-one streaming app some VTubers prefer
- VSeeFace / Warudo: common 3D avatar apps
- Live2D Cubism: the editor side of making a Live2D model
- Blender: frequent choice for 3D modeling
Software changes; the pipeline does not: art → rig → track → encode → platform.
If a term is missing here, it is probably either brand-new slang or a proper noun (a specific talent or software version). The stable core is still the same: a virtual YouTuber is a person, a model, and a live performance glued together by tracking.
Next reads: what is a VTuber, how to become a VTuber, and the software guide.