SpeechSuper provides cloud‑based APIs and offline‑ready SDKs that use deep‑learning models to evaluate spoken audio for pronunciation, fluency, grammar, and vocabulary across eight languages. The services deliver detailed scores at phoneme, word, sentence, and paragraph levels for scripted, semi‑scripted, and unscripted speech, enabling language learning apps and edtech platforms to offer real‑time feedback without building their own speech analysis infrastructure.
Funding
Funding not disclosed
Founders
Product
Problem
Language learners and educators lack scalable tools to automatically evaluate pronunciation, fluency, grammar, and vocabulary across multiple languages, making it difficult to provide immediate, detailed feedback in language learning applications.
Solution
SpeechSuper offers cloud‑based APIs and offline‑ready SDKs that analyze spoken audio using deep‑learning models. The services support scripted, semi‑scripted, and unscripted speech assessments for eight languages, delivering scores at the overall, phoneme, word, sentence, and paragraph levels. Users can submit audio recordings of single words, short texts (2–200 characters), or longer passages, and receive detailed metrics such as phoneme accuracy, stress detection, fluency, and IELTS‑style speaking scores. The platform provides per‑request pricing with unlimited request capacity, and the SDK enables integration into iOS and Android apps while preserving data privacy. This allows language learning platforms, edtech products, and assessment tools to embed real‑time pronunciation feedback without building their own speech‑analysis infrastructure.
Target Audience
Primary customers are language learning app developers, online tutoring platforms, and educational institutions that need automated speech assessment for learners of English and other major languages.
Features
- API endpoints for scripted pronunciation assessment (word, sentence, paragraph) across English, Mandarin, German, French, Spanish, Korean, Japanese, and Russian
- Semi‑scripted and unscripted speech assessment APIs with scores for fluency, grammar, vocabulary, and IELTS speaking criteria
- Phoneme‑level scoring, stress detection, and mispronunciation diagnostics with visualized results
- Offline‑ready SDKs for iOS and Android supporting single‑word, short‑text, and long‑text evaluations
- Unlimited request handling and per‑request pricing model starting at $0.0034 per request
- Data privacy safeguards with on‑premise SDK deployment and secure cloud transmission
- Comprehensive documentation and demo interfaces for each language and assessment type