Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
Wraps the Azure AI Voice Live SDK to enable real-time bidirectional voice applications in Python over WebSocket. Handles audio streaming, voice activity detection, transcription, and function calling in a single async API. Supports voice assistants, speech-to-speech translation, and voice-driven avatars.
Abstracts raw WebSocket management, audio encoding, and VAD logic so application code focuses on conversation flow rather than protocol details.
When to use it
- Voice assistant that listens via microphone and responds with spoken audio
- Real-time speech-to-speech translation during a live call
- Hands-free command interface for a Python desktop or server app
- Voice-enabled customer support bot with interrupt and turn detection
- Voice-driven avatar that reacts to live audio input
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: Build a Python voice assistant that listens to microphone input and responds with audio
- 1Connect to Azure Cognitive Services endpoint using DefaultAzureCredential
- 2Configure session with instructions, voice selection, and Server VAD settings
- 3Stream base64-encoded PCM16 audio chunks to conn.input_audio_buffer
- 4Listen for response.audio.delta events and decode audio bytes for playback
- 5Handle response.function_call_arguments.done to execute tools and return results to the conversation
An async Python app with real-time bidirectional voice conversation, automatic turn detection, and optional tool/function calling
Requirements
Accounts, API keys, or tools you or your AI assistant may need to set up while using this skill.