This paper presents the Aix-MARSEC project, an evolutive database of spoken British English. Specific details are given about the grapheme-phoneme conversion from the orthographic transcripts, the optimisation by elision rules of the phonetic transcription, the automatic phoneme-level alignment and automatic higher level treatment (syllables, subsyllabic structure, rhythmic units and MOMEL-INTSINT intonation coding). Integration of users' contributions will be within the general framework of GNU GPL licensing. Preliminary (pragmatic and prosodic) studies are presented in the final part of the paper.