If you are using Matlab and SPM, you can use the handy XJView software. Not only it lets you label the coordinates, but, if you select a cluster of voxels, it will give you a summary report of how many voxels belong to each area, using both the AAL and the Brodmann labels.
The Talairach demon works well (www.talairach.org) however you must remember to convert the coordinates to MNI to Talairach space (for example using icbm2tal in matlab, you can find it at www.brainmap.org/icbm2tal/)
The anatomy toolbox in SPM is very useful as is the already suggeted XJview and MRICROl. In my opinion, I would not recmommed Taliarach demon, as Taliarach coordinates are not very informative and are not the best way to localize to sulcal and gyral anatomy. I still tend to refer back to a classic atlas such as Duvernoy(1999) The Human Brain:
Surface, Three-dimensional Sectional Anatomywith MRI, and Vascularization. Although I appreciate it may not always be available though.
I used Matlab and SPM as Andrea Stocco from the University of Seattle did, but all other atlas suggested by other colleagues (See ALL ANSWERS) work very well.
Ciao Giorgia, I strongly agree with Matthew about the classic atlas Duvernoy (1999) or rather you can simply input your MNI coordinates in the MRIcron template (although I would say this is not very precise).