In games, we must project 3D points onto a 2D screen. This is done by multiplying local-space vertices by a series of matrices that move them through different coordinate spaces.
Each graphics API defines its own canonical view volume (CVV). Anything inside this volume is rendered; anything outside is clipped. The CVV differs across APIs.
OpenGL expects coordinates in the range -1 to 1, so all world-space positions must be transformed into clip space, which uses this range. It is called clip space because any vertex outside these limits is clipped. Let us look at how world-space vertices are converted into clip space.
Canonical View Volume (CVV)