Current web-based learning environments (such as WebCT or LearningSpace) are better at presenting the content than assessing the learning. Indeed, these environments provide sophisticated access to learning content by means of images, videos, sounds, hypertexts, glossaries, etc. However, their main weakness lies in the assessment part. Either the tests are based on rigid technologies such as quiz or multiple-choice questions (MCQ), or a teacher is required to manage forums or asynchronous e-mail exchanges. In this article we present Apex, a module of a web-based learning environment which is able to assess student knowledge based on the content of free texts. The goal of this module is not just to grade student productions but rather to engage students in an iterative process of writing/assessing at a distance. In addition to the preparation of their courses, teachers are required to partition them, to design exam questions, and to select the parts of the course that cover these questions. All courses, questions and student essays are managed by a database which is part of Apex. The assessment relies on Latent Semantic Analysis, a tool which is used to represent the meaning of words as vectors in a high-dimensional space. By comparing an essay and the text of a given course on a semantic basis, our system can measure how well the essay matches the course. Various assessments are presented to the student regarding the content, the outline and the coherence of the essay. These assessments provide a more authentic feedback than those