ST-1 Speak & Tweak is an emulation of the classic SP0256-AL2 speech chip that was a popular add-on for many home computers in the early 80s.
Speak & Tweak produces sounds using a crude model of the human vocal tract with 6 cascaded digital filters, white noise generator (breath) and tone generator. It contains 59 allophones which when concatenated correctly, can create most English words.
The result is a monotonic, robotic sounding voice that can be tuned precisely to any MIDI note, and is recognisable by anyone who's heard a speech chip from the early days of computing.
Version 1.1 also adds an audio input to allow you experiment with audio sources other than the built in pulse generator.