Recommender systems are software solutions that provide a list of suggestions which contains elements that can be of preference for each user. The use of these systems in multimedia content publish platforms facilitates the search of audiovisual content. The objective of this research is to analyze the techniques of collaborative filtering recommendation based on using the benefits of a large user community. In this research is carried out the description of the process flow for generating recommendations. Collaborative filtering algorithms based on users and items were analyzed and the evaluated algorithms with best results in the data set of the platform were selected. The main problems of the selected collaborative filtering technique, such as the problem of the new items, were also analyzed and solutions were proposed. The developed system was encapsulated in a module for VideoWeb 1.0 platform that uses the Drupal CMS in version 6. The results were evaluated using the mean absolute error method and presented using a range of 50 to 200 neighbors. The integration of the recommendations module to the platform provides an increase in the personalization of the multimedia content posted in order to satisfy each user preferences. This module offers an increase in the reliability of the users and minimizes the research time of the multimedia content.