Offline Recognition

SER and SEC inference on a full spoken utterance.

02

Audio to recognize

Idle

Record or upload a clip, then recognize it after recording ends.

Click to start

03

Recognition results

Result will appear here.
Recent results
  • No sessions yet.
SER labels: Neutral, Happy, Sad, Angry, Fear, Disgust, Surprise, Other/Complex. SEC returns a free-form description of emotion cues — prosody, tempo, voice quality, and other non-lexical signals.