v2.2.0

What's New

• Voice Streaming: Real-time voice analysis over WebSocket with live transcription, safety alerts, and session summaries
• Credits Tracking: Added credits_used field to all result models (BullyingResult, GroomingResult, UnsafeResult, etc.)
• Documentation: Added voice streaming and credits tracking documentation to README

Voice Streaming

session = client.voice_stream(
    config=VoiceStreamConfig(interval_seconds=10, analysis_types=["bullying", "unsafe"]),
    handlers=VoiceStreamHandlers(
        on_transcription=lambda e: print(f"Transcript: {e.text}"),
        on_alert=lambda e: print(f"Alert: {e.category} ({e.severity})"),
    ),
)
await session.connect()
await session.send_audio(audio_bytes)
summary = await session.end()

Install with voice support: pip install tuteliq[voice]