Voice recognition

Definition and details by Wikipedia

Free or open source

Sphinx
Julius, so far only with a Japanese language model and acoustic model. Used by Vestec for their Asterisk solution. See also here on asterisk forge.
ISIP
HTK - has distribution restrictions
IBM ViaVoice,see also http://volker.dnsalias.net/linux/speechrec.html, also available as Google cache or Wayback Machine. ViaVoice has software for both desktop and telephony applications.
http://perlbox.org/: Open source voice recognition for desktop steering
http://freespeech.sourceforge.net/FreeSpeech/html/
ZOIP: Text adventure ZORK for Asterisk (with text-to-speech and voice recognition)

LumenVox - amongst others available for Asterisk 1.4 (a LumenVox patch for 1.2 is also available)
Nuance (formerly ScanSoft): Professional solution, commercial TTS and ASR software often used for telephony applications
Loquendo: Professional solution
Acculab - Built-in to commercial DSP-based telephony cards
Lernout & Hauspie - Commercial software often used for telephony and embedded applications
Microsoft Speech Server - Does ASR and TTS with support for telephony applications
Other vendors: IBM, Fonix Speech, Sensory Inc., Verbio, Voice Signal Technologies, Vorero, Utopy
Asterisk cmd ASR - professional, multilanguage voice recognition for Asterisk based on ASP model
VoiceXML platforms typically offer TTS and ASR as part of their ASP, software, or hardware appliance solutions
Verbio ASR 8.0 - available for Asterisk in Spanish, Portuguese, English, French, Catalonian, Euskera, Gallician, Brazilian and differents kinds of American Spanish
Vocapia VoxSigma speech-to-text software suite