ELSA Speak

Feedback is phoneme-level rather than word-level, so a student learns which specific sound failed rather than just that a word was wrong. It gives learners unlimited low-stakes repetition without the face-threat of mispronouncing in front of classmates, and covers stress and intonation, not only segmentals.

Giving students pronunciation feedback they can practice against between class meetings, targeting the specific phonemes their first language does not contain. Assigning self-recorded minimal-pair drills students repeat privately until the score improves.

It is benchmarked against a narrow General American standard and will mark intelligible, perfectly acceptable accented English as wrong, which can damage the confidence of students who do not need to sound American to be understood. The recognizer itself misfires on heavily accented speech, producing feedback that is simply incorrect. Pricing is opaque and regionalized — verify in-app before recommending it to a class.