Give OpenClaw a Voice

Point OpenClaw at ListenHub and your assistant starts talking: replies are delivered as voice messages on platforms that support them, using ListenHub's expressive multilingual voices.

Requirements

• OpenClaw v2026.3.7 or later
• A ListenHub API key
• A ListenHub account with available credits — see Credits & Pricing

Setup

Add the following to your openclaw.json under messages:

{
  "messages": {
    "tts": {
      "auto": "always",
      "provider": "openai",
      "openai": {
        "baseUrl": "https://api.marswave.ai/openapi/v1",
        "apiKey": "$LISTENHUB_API_KEY",
        "model": "flowtts",
        "voice": "CN-Man-Beijing-V2"
      }
    }
  }
}

Replace $LISTENHUB_API_KEY with an API key from your settings. Replace the voice value with any speaker ID from the Speakers API. flowtts is ListenHub's TTS model name — leave it as-is.

Test with curl

Before wiring OpenClaw to ListenHub, verify your API key works:

curl -X POST https://api.marswave.ai/openapi/v1/audio/speech \
    -H "Authorization: Bearer $LISTENHUB_API_KEY" \
    -H "Content-Type: application/json" \
    -d '{
      "input": "Hello! This is ListenHub text to speech.",
      "voice": "EN-Woman-General-01",
      "response_format": "mp3",
      "model": "flowtts"
    }' \
    --output output.mp3

On success, output.mp3 is saved to the current directory. Swap voice for any ID from the Speakers API.

auto Modes

The auto field controls when TTS fires on outbound replies:

"inbound" pairs especially well with Telegram — voice in, voice out.

Available Voices

ListenHub provides natural-sounding voices across multiple languages. A few starting points:

Browse all available voices with the Speakers API.

Platform Support

Voice messages are delivered natively on these platforms: