In this paper we describe the development of two new text-to-speech reading modes for use in a computerised reading tutor for dyslexic children: a phoneme spelling mode which spells words phoneme by phoneme and a syllable mode, which allows speech to be synthesised as either isolated or lengthened syllables. These modes are used for modelling (i.e., to give a good example) and for giving feedback to the child. They are non-existent in most of the current reading tutors but will improve the flexibility of the tutor and allow better therapy.