Due to the paradigm shift from mass production (high volume, low variety) towards mass customization (high volume, high variety) today’s manufacturing systems must adapt to be more flexible. One approach for solving this is with the concept of Plug & Produce, which can be realized using a Multi-Agent System (MAS). In a MAS, the logical control is distributed onto individual, autonomous agents, which are digital representations of physical objects (parts and resources) in the manufacturing system. In these types of solutions agents representing robots can request paths for moving from A to B. These paths need to be generated automatically without human intervention to maximise flexibility. This thesis work aimed at implementing an automatic path planning service in a MAS.The work resulted in a system design for a path planner service that was successfully implemented and evaluated in a simplified multi-agent manufacturing scenario. The evaluation shows that it is possible to implement a path planner as an agent, the results include a comprehensive agent configuration in the system used. The solution is successful in generating a path upon request from an agent within the multi-agent system, without any need for human intervention once the system has been started. Suggestions for future work include improvements to the path planning algorithm used, tests on a real manufacturing cell and the development of a digital twin