openclaw - 💡(How to fix) Fix Feature request: per-agent TTS voice configuration [1 participants]

TL;DR

To assign different TTS voices to individual agents/bot accounts, modify the configuration to include agent-specific TTS settings, either at the channel or agent level.

Guidance

• Introduce agent-specific TTS configuration, as proposed in the channels or agents configuration shapes, to allow distinct voices for each bot account.
• Update the messages.tts configuration to reference the agent-specific TTS settings, ensuring that each bot account uses its assigned voice.
• Verify the configuration change by testing each bot account's TTS output to confirm that the correct voice is being used.
• Consider implementing a fallback or default voice setting for cases where an agent-specific voice is not defined.

Example

{
  "agents": {
    "list": [
      {
        "id": "iris",
        "tts": {
          "providers": {
            "openai": { "voice": "shimmer" }
          }
        }
      },
      {
        "id": "hermes",
        "tts": {
          "providers": {
            "openai": { "voice": "onyx" }
          }
        }
      }
    ]
  }
}

Notes

The proposed configuration changes assume that the OpenClaw instance is capable of handling agent-specific TTS settings. If this is not the case, additional development or modifications to the OpenClaw codebase may be necessary.

Recommendation

Apply a workaround by introducing agent-specific TTS configuration, as this allows for distinct voices for each bot account without requiring a version upgrade. This approach provides a flexible solution for managing multiple bot accounts with unique voices.