This research presents CoPi, a heuristic constraint solver designed for end-to-end scheduling of real-time task pipelines on multiprocessor systems. CoPi derives task execution times and periods while satisfying constraints on end-to-end delay, data loss, and schedulability. By converting pipelines into asynchronous, independent periodic tasks, the approach enables tasks from multiple pipelines to be mapped across processor cores while supporting dynamically appearing workloads and minimizing task migrations. Simulations show that CoPi achieves a comparable pipeline acceptance ratio to traditional optimization solvers while delivering significantly faster runtime performance.