Knots are cutpoints that defines different regions (or partitions) for a variable.
In each regions, a fitting must occurs.
The definition of different regions is a way to stay local in the fitting process.
One strategy is to decide K, the number of knots, and then place them at appropriate quantiles of the observed X.
Just spread them out uniformly, having roughly the same amount of data in each of the regions created by the knots.
The number of knots results in the number of parameters (see the formula)