01
Collect the recordings
Support calls, voicemails, user interviews, and post-call surveys — any prerecorded English audio you already keep.
Voice of customer
Voice-of-customer analytics from the voice itself
Most voice-of-customer programs stop at survey text. Run the actual recordings through the Speech API and you get voice of customer sentiment analysis from the audio — emotion and speaking style scored per sentence — turning support calls, voicemails, and interviews into structured voice-of-customer analytics you can trend, segment, and route.
One file / one result
Recorded feedback, scored for emotion and tone
Honestly, I was ready to cancel — but the way your team handled it completely turned me around.
relieved 0.74sincere 0.63hopeful 0.41
Unified analysis request
curl https://speech-api.oruk.ai/v1/audio/analysis \
-H "Authorization: Bearer $ORUK_API_KEY" \
-H "X-Request-ID: $(uuidgen)" \
-F "model=oruk-resonance" \
-F "file=@recording.wav"Workflow
02
Score emotion and style
Each file returns a transcript plus per-sentence emotion and speaking-style labels, so tone is measured from the voice, not inferred from text.
03
Trend and route
Aggregate the labels into voice-of-customer analytics — trend sentiment over time, segment by cohort, and route the moments that need a human.
Useful applications
- Voice of customer sentiment analysis from support calls and voicemails
- Voice-of-customer analytics: trend emotion over time and by segment
- Surface at-risk or delighted customers from how they sound
- Attach acoustic tone to interview and survey transcripts
Deployment safeguards
- Treat emotion labels as acoustic measurements, not statements of intent
- Validate on the audio conditions, languages, and accents in your program
- Keep multiple labels and uncertainty rather than one definitive state
- Disclose recording and analysis according to applicable consent rules