I'm pleased to announce version 0.5 of the Python speech recognition module, speech.py. What is speech.py? =============== speech.py provides a clean and simple interface to the Microsoft Speech Kit, allowing your Windows Python program to speak out loud and to recognize spoken input. Homepage: http://pyspeech.googlecode.com/ PyPI: http://pypi.python.org/pypi/speech/ Installation: "easy_install speech" (you may also need one or two programs that are available via the homepage) What's new in version 0.5? =================== The most important addition is support for synchronous speech recognition, allowing you to detect speech much as raw_input() detects typing. For example, the following code detects speech and repeats what was heard: import speech answer = speech.input("Say something.") speech.say("You said: %s" % answer) answer = speech.input("Say a number less than three.", ["Zero", "One", "Two"]) speech.say("You picked %s" % answer) (This is in addition to the existing asynchronous API, allowing you to register callbacks to execute whenever certain spoken text is heard.) This release also represents a stabilization of the API and documentation -- thank you to the users to worked with early versions of the module as the API was refined. This will likely be the last release announced before 1.0, when support for grammar state machines is added. Example application: MusicButler ======================== An example application is available at http://musicbutler.googlecode.com -- a robot that learns your MP3 collection and then lets you control your stereo via voice interaction.. MusicButler is in alpha and doesn't warrant its own announcement yet; but it is a good example of a complex speech application.