This study aimed to develop a secure, peer to peer-based search intelligence system. The system mentioned in this paper which using the desktop search technology based on metadata can reduce the need of the central server by introducing the parallel computing. We focus on solving the problems including privacy, security and incompletion by introducing a new architecture. The infrastructure provided by the search engine can be utilized to simplify the users' action as well. The system first finishes metadata retrieval, then submits those metadata to the system to let other nodes know these data. The central server helps the nodes to communicate with each other, schedules the nodes actions and track their states. The result shows that the consuming work is moving to the client side, leaving only some analysis work to the server.