String field theory is a proposal to define string theory in such a way that the background independence is respected. String field theory can be understood as a quantum field theory with infinitely many fields which are unified into one master "string field". This string field, roughly speaking, assigns an operator to every one-dimensional curve (string) in spacetime, much like a quantum field assigns an operator to each zero-dimensional point in spacetime.

String field theory did not turn out to be helpful in the second superstring revolution because this revolution has revealed that other objects such as branes are as fundamental as the strings themselves. String field theory is based on the assumption that the strings are the fundamental objects, and it makes it more difficult (or impossible) to understand dualities within its framework.

There are several versions of string field theory - for example the boundary string field theory or the cubic (Chern-Simons-like) string field theory constructed by Edward Witten. In the late 1990s, both of them turned out to be very useful to understand tachyon condensation.