discord-ai-call

Multi-Voice AI Discord Bot

A Discord bot that provides multiple AI personalities with different voices for natural group conversations. Uses Gladia for speech-to-text, GPT-5 nano for AI responses, and Murf AI for text-to-speech.

Features

Multiple AI Personalities

• Maverick (28, en-US-maverick) - Calm, composed, and optimistic
• Ruby (24, en-UK-ruby) - Charming, energetic, and enthusiastic
• Charles (26, en-US-charles) - Witty, analytical, and curious

Natural Conversations

• Context-aware responses that reference previous messages
• Direct addressing: “Hey Charles, what do you think?”
• Smart voice selection based on conversation flow
• Conversation memory

Advanced Audio

• Real-time speech-to-text via Gladia WebSocket
• High-quality TTS via Murf AI WebSocket
• Multi-language support

Setup

1. Install Dependencies

npm install

2. Get API Keys

You’ll need API keys from:

• Discord: Create a bot at Discord Developer Portal
• Gladia: Get your API key from Gladia Dashboard
• OpenAI: Get your API key from OpenAI Dashboard
• Murf: Get your API key from Murf API Dashboard

3. Configure Environment Variables

Update the .env file with your API keys:

# Discord Bot Configuration
DISCORD_TOKEN=your_discord_bot_token
CLIENT_ID=your_discord_application_id_here

# API Keys
GLADIA_API_KEY=your_gladia_api_key
OPENAI_API_KEY=your_openai_api_key
MURF_API_KEY=your_murf_api_key_here

4. Discord Bot Permissions

Make sure your Discord bot has these permissions:

• Connect to voice channels
• Speak in voice channels
• Use Voice Activity

Usage

1. Start the bot:

npm start

1. The bot will:
   • Join the specified voice channel
   • Listen for users speaking
   • Transcribe speech to text
   • Generate AI responses
   • Convert responses to speech
   • Play audio back in the channel

Technical Details

Audio Format Requirements

• Input: Discord provides 48kHz, Opus-encoded audio
• Processing: Converted to PCM for Gladia transcription
• Output: Murf generates 48kHz WAV for Discord compatibility

Architecture

Discord Voice → Prism-Media → Gladia STT → OpenAI → Murf TTS → Discord Audio Player

Files

• index.js - Main Discord bot logic and voice handling
• gladia.js - Speech-to-text integration with Gladia API
• openai.js - AI response generation with OpenAI
• murf-streaming.js - Text-to-speech integration with Murf AI

Voice Customization

You can customize the TTS voice in murf.js by changing the voiceId parameter:

const voiceId = "en-US-ken"; // Change to your preferred voice