The Net4Voice project aims to increase the quality of learning opportunities promoting the adoption of barrier-free learning environments and the development of innovative methodologies which use speech recognition (SR) technologies. SR technologies can automatically transform a lecturer’s speech into digital text in real-time, generating an electronic transcription of the lesson or conference ready to be printed or delivered through different devices and channels. This paper intends to disseminate information and results obtained so far through the Net4Voice project, financed by the European Commission under the Lifelong Learning Programme. In particular the objectives of this article are to present the project's objectives and the activities implemented to achieve them; describe the learning methodology developed and show the first outcomes of ongoing tests. Net4Voice started in December 2007 and ends in November 2009. The project involves three universities and two high schools, where students with different kinds of auditory or motor disabilities can particularly benefit from the use of speech recognition technologies to follow lessons in an accessible learning environment without additional intermediary support. Second language learners also benefit from these technologies, as once the class has finished they are able to access the multimedia transcriptions available and read or listen the lessons again. Net4Voice aims at defining a clear scheme to assess the impact of speech recognition technologies in education, by experimenting it in at least three different learning contexts: university, school and adult education classrooms. Moreover, the project monitors the effectiveness of a new learning methodology, a pedagogical support system, in order to use technology at its best, exploiting its potential, adapting it to students’ conditions and needs and combining software with personal teaching experience and methods.