This paper presents an approach to generate audio based learning material dynamically from Wikipedia articles for m-Learning and ubiquitous access. It introduces the so called ’Pediaphon’, an speech interface to the free Wikipedia online encyclopedia as an example application for ’microlearning’. The effective generation and the deployment of the audio data to the user via podcast or progressive download (pseudo streaming) are covered. A convenient cell phone interface to the Wikipedia content, which is usable with every mobile phone will be introduced.