Cutting-edge speech recognition engine specially developed for ALL* applications.
Robustness to speaker variability and elimination of ambient noise through state-of-the-art acoustic models.
Over a decade of research and development by a large team of highly experienced researchers, engineers, linguists, phoneticians, and pedagogical experts.
Statistical analysis and categorization of errors provide valuable and accurate pedagogical feedback on frequent pronunciation difficulties and practical guidance on how to correct them.
Highly accurate sound recognition which allows detection of even the slightest deviations from native English target sounds.
Unique handling of connected speech, such as variations and reductions, prevents false feedback during dialogue practice.
Unique methodology achieves a high correlation with pedagogical experts’ assessments.
Engine performance is highly optimised through analysis of typical mother tongue errors.
Voice-activated selection of multiple choices in dialogs.
Advanced speaker adaptation improves recognition accuracy.
Complete our short, free test to get an indication of your current level of English.