10.6x
Faster scoring
Instant pronunciation reports after STT completes.
Enterprise-grade AI assessment
The new standard for AI pronunciation assessment.
Measure pronunciation accuracy with deterministic, auditable scoring built for education, language learning, and assessment platforms.
Live pronunciation pipeline
deterministicAudio waveform
Speech to text
Phoneme alignment
Pronunciation score
Learner feedback
Assessment report
92
Excellent clarity
Pronunciation94%
Fluency89%
Rhythm91%
Try live demo
Understand DUNA in under 30 seconds.
Pick a bundled recording, run the pipeline, hear the audio, and see a complete pronunciation report without uploading anything.
Excellent delivery
Expected: The quick brown fox jumps over the lazy dog
Upload
Speech Recognition
Transcript
Phoneme Alignment
Scoring
Feedback
Assessment Report
Click Run Demo to animate the complete assessment pipeline.
Trusted by modern education platforms
Lexora
Speakly
Edvo
Lingua
TutorOS
Auralab
37%
More actionable learner feedback
Word-level mistakes, timestamps, and practical correction tips.
4.8x
Auditable assessments
Deterministic math instead of opaque generated scoring.
Product
Designed for high-stakes pronunciation assessment.
A guided assessment workspace with consent, passage setup, recording or upload, and a premium learner report.
1Consent
2Passage
3Audio
4Score
1
Consent
Required before recording or uploading. Audio is processed temporarily and never persisted by the app.
DPDP consent
Temporary audio processing only
Your audio is processed temporarily and deleted within 24 hours. No data is stored permanently. If a cloud speech-to-text provider is configured, your audio is transmitted to that provider only to generate the transcript used for this score. Consent is required before the microphone can start recording.
Passage
Paste the learner prompt. The scoring engine compares this text against the real STT transcript.
23 words. Short recordings are okay; missing words will be scored as part of the assessment.
3
Audio
Record with the browser mic for the fastest path. Upload is secondary.
Record in browser
Consent first, then click Start speaking and choose the microphone you want to use.
Timer
0:00
Noise level
Standby
Real-time transcript
Select an external microphone if needed, then start recording to generate a transcript and pronunciation report.