Spatial messaging is a direct extension to text and other multi-mediamessaging services that have become highly popular with the currentpervasiveness of mobile communication. It offers benefits especiallyto mobile computing, providing localised and therefore potentiallymore appropriate delivery of nearly arbitrary content. Location isone of the most interesting attributes that can be added to messagesin current applications, including gaming, social networking, oradvertising services. However, location is also highly critical interms of privacy. If a spatial messaging platform could collect thelocation traces of all its users, detailed profiling would be possible– and, considering commercial value of such profiles, likely. Inthis paper, we present emphAir-Writing, an approach to spatialmessaging that fully preserves user privacy while offering globalscalability, different client interface options, and flexibilityin terms of application areas. We contribute both an architectureand a specific implementation of an attribute based messaging platformwith special support for spatial messaging and rich clients for J2ME,Google Android, and Apple iPhone. The centralised client/server approachutilises groups for anonymous message retrieval and client cachingand filtering as well as randomised queries for obscuring traces.An initial user study with 20 users shows that the overall conceptis easily understandable and that it seems useful to end-users. Ananalysis of real-world and simulated location traces shows that userprivacy can be ensured, but with a trade-off between privacy protectionand consumed network resources.