Adaptive versions of multilevel splitting algorithms have become more and more popular in the recent years. They are used in many domains of physics like molecular dynamics, chemistry, networks, probability or even geophysics and fluid mechanics. They offer a fast and simple way for computing events with very small probabilities by performing successive selections and mutations on the system in a controlled and adaptive way. They are usually exponentially better than direct Monte-Carlo approaches. After presenting the algorithm in details, we will show various applications for metastable systems.