We present a design environment which allows the generation, modification, and visual speech animation of 3D cartoon-like faces - Tinky. Our underlying face model is not based on a set of independent parameters that control specific abstract muscle emulations but is directed by a set of objects representing the elements of the face. In order to provide an easy to use authoring system producing standardized visemes, expressions and animations of web oriented talking faces, we adopt an object oriented approach for face design and combine it with the authoring qualities of VRML and the standardized set of facial animation parameters defined by MPEG-4 SNHC.