IPSVM is a machine learning algorithm for accurate detection of interplanetary shocks in spacecraft data, which has emerged as a proceeding to another project we were involved into - IPShocks. Manual detection of shocks, which are relatively small phenomena on temporal scale (minutes), in long time series (decades) is an extremely tedious and time consuming labour. IPSVM automates this task completely thus saving lots of resources. The algorithm is at least twice as precise as all existing shock detection solutions that we are aware of. Currently IPShocks database is filled with new shocks with the help of IPSVM.