Google recently allowed text to be spoken using the Text-To-Speech API . Now it is also possible to reverse the conversion of speech into text.
Using speechapi.com it is possible to build interesting speech web resources that include two-way text-to-speech and speech-to-text (speech-to-text as and text-to-speech) conversions.
This was made possible through the use of various technologies: flash is used to access the microphone and streaming audio, Red5 - as a versatile open-source media server.
( Link to the original )
An example of using the API: