eSpeak is a brilliant opensource text to speech project. It's the default TTS software on many Linux distributations, but there's a Mac version available and it works great. There is also a project based on eSpeak called Mbrola which adds "singing" functionality.
Fwiw, I really don't see Apple suing you if you use the built-in TTS they provide.
Dunno if this is exactly what you want, but GarageBand has options for vocal tracks (Mouse Voice, Deeper Voice, Helium Breath, etc.) so you could speak the text yourself and then alter it. It is then usable commercially, as the GarageBand software license agreement says:
GarageBand Software. You may use the Apple and third party audio loop content (Audio Content), contained in or otherwise included with the Apple Software, on a royalty-free basis, to create your own original music compositions or audio projects. You may broadcast and/or distribute your own music compositions or audio projects that were created using the Audio Content, however, individual audio loops may not be commercially or otherwise distributed on a standalone basis, nor may they be repackaged in whole or in part as audio samples, sound effects or music beds.