The Behavioral Signals API Developer Hub
The Behavioral Signals API analyzes speech for how it was spoken. Send an audio file or stream it live, and get back results per utterance: who spoke, what they said, and how they said it: emotion, positivity, strength, engagement, hesitation and speaking rate, along with speaker gender and age.
The same API detects synthetic speech. The deepfake endpoints return a bonafide or spoofed verdict per utterance and, where identifiable, the generator most likely behind it. Video is supported for visual deepfake detection.
One set of credentials, one response shape, three jobs chosen by the endpoint you post to. Batch over REST, or real-time over gRPC with predictions arriving as the audio does.