Frontier voice-native models for transcription, deepfake detection, PII/PHI redaction, music & speech detection, emotions, and behavior signals, available as REST and streaming APIs.