Moodzle detects the emotion in human voice samples. The EmotionAPI allows user to upload an audio sample then a data structure containing underlying emotion, as well as speech-to-text, is returned. Free sandbox accounts are available, as are paid production accounts. Accuracy levels above reported as 99% for the training set, over 91% for non-training data, including male and female voices, noisy environments, smartphone microphone and high quality microphone settings. An account is required with service, the API uses REST calls.