I'm working on an application that would require continuous speech recognition. The app will listen for only a few defined keywords, but once the listener gathers the users voice input, it will need to keep listening for more keywords until the user stops the listener.
The best solution I've found is the Nuance speech recognition library, but this isn't perfect and it could get very expensive. Has anyone else been able to get this type of functionality working?
You can try the Oxford API of Mircosoft
This can give you continuous real time speech recognition, and then you can look for special keywords. Maybe it's an overkill in your case so be sure to check more options !