version: "1.0.0" required_env: [OPENAI_API_KEY] platforms: [linux, macos, windows] setup_notes:

• Configure an active OpenAI model before using STT or TTS routes.
• Use voice mode for short operator interactions, not large document dumps.

Voice Ops

Use voice features when an operator needs speech input, spoken output, or hands-busy interaction from WebChat.

• Keep spoken responses shorter than typical text responses.
• Fall back to text if the active OpenAI voice model is missing or quota-limited.
• Confirm the target language when transcribing mixed-language audio.