Pronunciation feedback — live demo
Pronunciation, accuracy, fluency, completeness, prosody — draft scores. In an exam, the marker reviews, overrides, signs off.
Practice Setup
loading…
You are interacting with an AI system. Clicking the microphone gives NUADU temporary access to your device's microphone while you speak. You can stop at any time.
How your audio is processed
How your audio is processed: your speech is streamed from your browser over HTTPS directly to Microsoft Azure AI Speech in the EU (Sweden Central), using a temporary access token valid for about ten minutes; it does not pass through NUADU's servers. Azure analyses it in real time and returns the recognised text and, in the pronunciation demo, pronunciation scores. In the speaking demo the recognised text is then scored by Azure OpenAI in the same region. NUADU does not store your recording; we receive only the text and the scores. Microsoft processes the audio as our processor under the Microsoft Products and Services DPA and does not use it to train models.
Controller and contact: NUADU Holding Pte. Ltd., Singapore — privacy@nuadu.com. Legal basis: legitimate interests (Art. 6(1)(f) GDPR) in providing the analysis you requested; an assessment is available on request.
AI-generated results: the scores are produced by automated processing, not by a human examiner. They are a demonstration, may vary with microphone quality and background noise, and must not be used as the sole basis for any educational or professional decision. You may contest a result or ask about the processing at privacy@nuadu.com.
Details of the models and data handling: AI Transparency Notice · Privacy Policy.