01
Upload completed audio
Send a supported file after the call. API v1 is prerecorded speech analytics; it does not claim real-time intervention.
Contact-center intelligence
Contact center analytics, from the audio of every call
Turn completed service calls into contact center analytics: a transcript plus multilabel emotion, speaking style, and time-local segments. Score a whole queue for tone, find the calls worth a careful human review, and compare patterns over time — without pretending a score is a person’s inner state.
One file / one result
Every call scored for emotion and speaking style
I have already explained this twice, and I need someone to resolve it today.
frustrated 0.91impatient 0.78energetic 0.42
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
Receive aligned outputs
Store the transcript, per-sentence emotion labels, segments, request ID, and usage record in your existing call-analytics or QA workflow.
03
Route human review
Feed the acoustic signal into coaching and quality review alongside operational context and qualified reviewer judgment.
Useful applications
- Prioritize a manageable review queue instead of sampling calls at random
- Score an entire queue for tone rather than the few percent QA can listen to
- Compare aggregate patterns across queues, teams, or time periods
- Locate the specific moments worth replaying in a coaching session
- Add acoustic context to transcript search, so a query can find how it was said
- Flag calls where the caller stayed frustrated through to the end
Deployment safeguards
- Do not use emotion labels as the sole employee-performance measure
- Validate on the microphones, languages, accents, and call types in production
- Retain uncertainty and multiple labels rather than one definitive state
- Disclose recording and analysis according to applicable consent rules