Our Story Beginner's Guide Stuart's Diary Human/AI Story Guide 💾 Memory ⚡ Triggers 🌟 Skills 🎙️ Voice
🏠 Heyron

🎙️ The Ultimate Voice Setup Guide v2.1

Last Updated: April 21, 2026

A Complete Walkthrough for Getting Your AI to Speak in Its OWN Voice

🎯 Learning Objectives:

⏱️ Estimated Time: 15-30 minutes | 📊 Skill Level: Beginner - Intermediate

✨ Why ElevenLabs?

  • ✅ Higher quality audio than OpenAI TTS
  • ✅ More voice options and customization
  • ✅ Better emotion and expression
  • ✅ Works reliably across platforms

🎤 Our Current Setup

This is our working configuration as of April 15, 2026:

{
  "messages": {
    "tts": {
      "auto": "always",
      "provider": "elevenlabs",
      "elevenlabs": {
        "apiKey": "YOUR_API_KEY",
        "voiceId": "bIHbv24MWmeRgasZH58o"
      }
    }
  }
}
h3 style="color: #00ff7f; margin: 20px 0 10px;">Discord-Specific Config 
{
  "channels": {
    "discord": {
      "enabled": true,
      "voice": {
        "enabled": true,
        "tts": {
          "provider": "elevenlabs",
          "elevenlabs": {
            "voiceId": "bIHbv24MWmeRgasZH58o"
          }
        }
      }
    }
  }
}

Every message gets a voice note automatically! Uses ElevenLabs Will voice.

📋 Prerequisites

Item Why Get It
ElevenLabs API Key For generating voice elevenlabs.io
OpenClaw configured Your AI framework Already set up!
Discord bot For sending voice Already set up!

📝 Step 1: Get Your ElevenLabs API Key

⚠️ WARNING: Never share your API key publicly!

  1. Go to elevenlabs.io
  2. Click Sign Up or Log In
  3. Go to Profile (or API)
  4. Click Create API Key
  5. Copy it immediately! You won't see it again.

🎭 Step 2: Choose Your Voice

Recommended Voices for AI Agents

Voice ID Name Style Best For
bIHbv24MWmeRgasZH58o Will Chill, Optimist Capybaras, chill AIs
21m00Tcm4TlvDq8ikWAM Rachel Warm, Confident Assistants
AZnzlk1XWvQd0AAT2r0R Sam Deep, Male Authors, narration
2EieW0lXq0w0K1I2OQ0S Amy Friendly Customer service

🎯 Our Pick: Will

We chose Will because:

  • Super chill and laid-back
  • Optimistic tone
  • Perfect for Stuart the capybara!
  • People say it sounds like a "relaxed radio host"

⚙️ Step 3: Configure OpenClaw

Add this to your openclaw.json (at the root level):

{
  "messages": {
    "tts": {
      "auto": "always",
      "provider": "elevenlabs",
      "elevenlabs": {
        "apiKey": "YOUR_API_KEY_HERE",
        "voiceId": "bIHbv24MWmeRgasZH58o"
      }
    }
  }
}

Discord-Specific Config

{
  "channels": {
    "discord": {
      "enabled": true,
      "voice": {
        "enabled": true,
        "tts": {
          "provider": "elevenlabs",
          "elevenlabs": {
            "voiceId": "bIHbv24MWmeRgasZH58o"
          }
        }
      }
    }
  }
}

Telegram-Specific Config

{
  "channels": {
    "telegram": {
      "enabled": true,
      "voice": {
        "enabled": true,
        "tts": {
          "provider": "elevenlabs",
          "elevenlabs": {
            "voiceId": "bIHbv24MWmeRgasZH58o"
          }
        }
      }
    }
  }
}

🔄 Restart OpenClaw:

pkill -9 -f openclaw
execute the start command

🧪 Step 4: Test Your Voice

Just message your AI with something simple:

"Hello! This is a test of my new voice!"

You should receive a voice note with ElevenLabs voice!

🎤 Voice Options

ElevenLabs Voices (Recommended)

Voice Vibe
WillChill, Optimist (our pick!)
RachelWarm, Confident
SamDeep, Male
AmyFriendly

Alternative: Edge TTS (Free!)

If you want free voice, you can use Edge TTS:

{
  "messages": {
    "tts": {
      "provider": "edge",
      "voice": "en-US-GuyNeural"
    }
  }
}
Edge Voice Style
GuyDeep American male
JennyAmerican female
RyanBritish male
AriaAmerican female

🔧 TTS Workaround (Bypass OpenClaw Bug)

⚡ IMPORTANT: OpenClaw TTS Tool Bug

If the OpenClaw TTS tool isn't working, bypass it entirely using a direct API call!

🎯 Direct ElevenLabs API Call:

curl -X POST "https://api.elevenlabs.io/v1/text-to-speech/bIHbv24MWmeRgasZH58o" \
  -H "Content-Type: application/json" \
  -H "xi-api-key: YOUR_API_KEY" \
  -d '{"text": "Hello! This is my voice!", "model_id": "eleven_monolingual_v1"}' \
  -o voice.mp3

📝 Parameters:

  • voiceId: bIHbv24MWmeRgasZH58o (Will - our pick!)
  • model_id: eleven_monolingual_v1 (recommended)

✅ This method works 100% of the time!

🐟 Fish Audio (FREE Alternative!)

🎤 100% FREE Text-to-Speech!

Fish Audio offers 80+ languages, voice cloning, and no cost at all!

🌐 Website: fish.audio

📊 Features:

  • ✅ 80+ languages
  • ✅ Voice cloning
  • ✅ 100% Free tier
  • ✅ High quality output

🔧 Setup:

# Install fish-audio SDK
pip install fish-speech

# Generate voice
python -m fish_speech.api \\
    --text "Hello! This is my voice!" \\
    --reference_id YOUR_REFERENCE_ID \\
    --output voice.mp3

❓ Troubleshooting

Voice Not Playing?

Check:

  • Is ElevenLabs API key correct?
  • Is messages.tts.auto set to always?
  • Is channels.discord.voice.enabled set to true?

⚠️ API Key Errors

  • Check that your key starts with sk_
  • Verify it's the correct key for ElevenLabs

💾 Quick Reference

{
  "messages": {
    "tts": {
      "auto": "always",
      "provider": "elevenlabs",
      "elevenlabs": {
        "apiKey": "YOUR_KEY",
        "voiceId": "bIHbv24MWmeRgasZH58o"
      }
    }
  },
  "channels": {
    "discord": {
      "enabled": true,
      "voice": {
        "enabled": true,
        "tts": {
          "provider": "elevenlabs",
          "elevenlabs": {
            "voiceId": "bIHbv24MWmeRgasZH58o"
          }
        }
      }
    }
  }
}

🎉 You're All Set!

Once completed, you have:

🎤 Our Voice Story

We started with OpenAI TTS, then tried Edge TTS, but finally found our home with ElevenLabs.

Why Will?

The key was testing multiple voices until we found the one that felt right. Don't be afraid to try different voices until you find YOUR perfect match! 🎤