Gestures enable users to control avatars to do more than navigate. A typical MU system will have a client interface with a user interface with buttons that can be pressed to tell the user's avatar to perform a gesture that the avatar is capable of.

Typical gestures include: nodding, shaking head, waving, smiling, crying, jumping

Avatars in VNet, for example, can have a structure like this:

PROTO Avatar [
  exposedField MFString gestures [ "aGesture", ... ]
  eventIn      SFBool   aGesture
  ...
] {
  
  # Geometry goes here.
  ...

 
  # Interpolators, Scripts, and ROUTES go here.

  # Should at least handle what to do with the aGesture eventIn
  ...
}

This prototype to the MU system exposes the gestures that the avatar is capable of so that clients can perform them when they are used by the user.

Several VRML MU systems support avatars in this format, including "VNet" and "Deep Matrix".