For the web server, the profile link databases constructed lead the in-link for new protein sequences cannot be got. Therefore, a hybrid version of PL-search for the web server is proposed. In the web server, when calculating the similarity of protein pairs, the first level of Jaccard distance is calculated by out-link and profile link and the calculate manner of second level of Jaccard distance is retained. The final ranking list is consisted by our search result and out-link instead of double-link. Furthermore, the parallel speed-up is implemented in webserver and stand-alone tool.