Pronunciation Assessment APIs
Deep learning-powered speech and pronunciation assessment of pronunciation, fluency, and grammar.
SpeechSuper API scores paragraphs, sentences, words, phonemes, and tones. It also scores pronunciation, fluency, and completeness for spoken sentences and paragraphs.
SpeechSuper API detects whether the user laid the stress on the right syllable. It is useful for users speaking non-syllabic languages as their first language.
SpeechSuper API diagnoses what phonemes users say so that products can give instructions for mispronunciation correction.
SpeechSuper API supports scoring for children, young adolescents, and adults; customizing pronunciation for specific words; returning scores at different grading scales and decimal places. The scoring strictness is flexible based on your needs.
We support word-level, sentence-level, and paragraph-level closed-end assessments. It’s useful for prompted read-aloud speaking activities and practice.
Semi-open-ended assessment needs multiple reference texts to give scores. It also supports scoring with key points. We’re now working towards an open-ended assessment API.
Support for Accents
SpeechSuper API supports American, British, and Australian English. It is not biased toward any accents and supports specifying a particular accent to score.
Phoneme Spelling Alignment
Reading paragraphs could be boring. SpeechSuper API’s real-time feedback can make the experience interactive and engaging. It can light up the words users just read to behave like Karaoke.
What We Served
Trusted and Used by
Available in English, Mandarin Chinese, German, French, Korean, Japanese, Russian and Spanish.
up to 20s
up to 90s
less than 200 words per request
up to 300s
less than 1K words per request
Use 1M requests, get 50K requests for free.
Use 5M requests, get 500K requests for free.