Understanding how musicians can coordinate their musical actions when they improvise together remains an important theoretical and empirical challenge. In this article, we suggest a broad theoretical framework, compatible with up-to-date research on joint action, which can account for coordination in collective improvisation, especially in the hard case of so-called collective free improvisation. This framework addresses the limits of an account of coordination in collective improvisation that relies only on low-level, emergent coordination mechanisms, and shows how these mechanisms can be combined with planned coordination mechanisms to explain how improvisers deal with some of the main coordination problems that typically arise in collectively improvised performances. As such, our framework allows for the formulation of new hypotheses that pave the way for further empirical investigations on collective improvisation and sheds light on collectively improvised behavior at large.