Skip to content

Free tool

Voice emotion analyzer

Press the button, speak, and watch your words appear with live emotion and speaking-style labels. Runs on the production oruk Speech API — audio is analyzed, never stored.

Try it live — tap and speak

What the analyzer returns

As you speak, the demo chunks your audio and sends it to POST /v1/audio/analysis, which returns the transcript plus calibrated multilabel scores. Emotion labels (15): happy, excited, hopeful, sad, worried, angry, frustrated, disappointed, scared, disgusted, surprised, embarrassed, proud, relieved, neutral. Speaking-style labels (16): energetic, passionate, irritated, warm, playful, sarcastic, deadpan, hesitant, confident, sincere, skeptical, tired, formal, casual, impatient, distracted.

Scores are calibrated on held-out audio, so a label is only attached when it crosses its measured threshold. Outputs are acoustic annotations of how speech sounds — not claims about a speaker’s inner state. See capabilities and scope.

FAQ

Is this voice emotion analyzer free?
Yes. The demo on this page runs against the live oruk Speech API at no cost and with no account. For your own applications, new API accounts include $5 in trial credit.
Is my voice recording stored?
No. Audio is captured in your browser, sent to the API over TLS, analyzed, and discarded once the response is returned. Nothing is retained or used for training.
How does the analyzer detect emotion from voice?
An acoustic model scores 15 emotion labels and 16 speaking-style labels directly from the audio signal — pitch, pace, energy, timbre — with thresholds calibrated on held-out audio. Labels describe how the speech sounds, not what the speaker inwardly feels.
What languages does it support?
The transcription and labels are built for English audio. Other languages may return output, but accuracy is not evaluated or supported in API v1.
Can I use this in my own product?
Yes — the demo calls the same POST /v1/audio/analysis endpoint you would integrate: one REST request with an audio file returns the transcript, emotion labels, style labels, and time-local segments.

Build this into your product

The same endpoint powers call analytics, meeting tools, and research pipelines. Create an account for $5 in trial credit — no card required.

Speech emotion recognition API Benchmarks Pricing