A speaker profile is acoustic data used by a speech recognizer to increase its recognition accuracy for a specific individual's voice and environment. Because these files are constantly updated during speech recognition, they can become corrupted or damaged. It is important to be able to restore the speaker profile or dynamically create a new one on demand to minimize application disruption. Use Chant ProfileKit to administer speaker profiles for more accurate speech recognition.