Software testing is a key factor on any software project; testing costs are significant in relation to development costs. Therefore, it is essential to select the most suitable testing techniques for a given project to find defects at the lower cost possible in the different testing levels. However, in several projects, testing practitioners do not have a deep understanding of the full array of techniques available, and they adopt the same techniques that were used in prior projects or any available technique without taking into consideration the attributes of each testing technique. Currently, there are researches oriented to support selection of software testing techniques; nevertheless, they are based on static catalogues, whose adaptation to any niche software application may be slow and expensive. In this work, we introduce a content-based recommender system that offer a ranking of software testing techniques based on a target project characterization and evaluation of testing techniques in similar projects. The repository of projects and techniques was completed through the collaborative effort of a community of practitioners. It has been found that the difference between recommendations of SoTesTeR and recommendations of a human expert are similar to the difference between recommendations of two different human experts.