AI-Digi-Ellie

AI-Digi-Ellie is an AI-powered companion. She can operate in two modes:

• Discord mode: Runs as a Discord bot when provided with a Discord token
• Console mode: Runs in console for local testing when no token is provided

Features

• AI-powered conversation capabilities
• Discord integration using the D++ library
• Text-to-Speech support using Azure Cognitive Services
• Speech-to-Text support using Whisper.cpp (local processing)
• Voice chat interaction in Discord
• Supports both Discord bot and console modes

Requirements

Build Requirements

• CMake 3.10 or higher
• C++20 compatible compiler
• Git (for cloning the repository)
• OpenSSL 3.0 or higher (for secure connections)
• Opus development libraries (for voice processing)

Dependencies

The project uses the following libraries (included as submodules):

• D++ - A lightweight C++ Discord library
• nlohmann/json - JSON for Modern C++
• spdlog - Fast C++ logging library
• cpp-httplib - A C++ HTTP/HTTPS client library
• whisper.cpp - C++ implementation of OpenAI's Whisper model

Configuration

The project uses environment variables for configuration. Set the following variables before running:

Required Environment Variables

• DIGI_ELLIE_DISCORD_TOKEN - Your Discord bot token
• DIGI_ELLIE_DEFAULT_CHANNEL_ID - Discord channel ID for boot messages

Optional Environment Variables

• DIGI_ELLIE_MODEL_NAME - Ollama model name (default: "mistral")

TTS Configuration (Azure)

To enable TTS functionality, set these environment variables:

• DIGI_ELLIE_AZURE_SPEECH_KEY - Your Azure Cognitive Services subscription key
• DIGI_ELLIE_AZURE_SPEECH_REGION - Azure region (default: "germanywestcentral")
• DIGI_ELLIE_AZURE_SPEECH_VOICE - Voice to use (default: "en-US-JennyNeural")
• DIGI_ELLIE_AZURE_SPEECH_APP_NAME - Application name for Azure requests

STT Configuration (Whisper)

The Whisper STT service runs as a separate process and can be configured with:

• DIGI_ELLIE_WHISPER_MODEL_NAME - Whisper model file name (default: "ggml-large-v3-turbo-q8_0.bin")
• DIGI_ELLIE_WHISPER_SERVICE_HOST - Host for the Whisper service (default: "0.0.0.0")
• DIGI_ELLIE_WHISPER_SERVICE_PORT - Port for the Whisper service (default: 8000)

Building the Project

Linux/macOS

1. Clone the repository with submodules:

git clone --recursive https://github.com/yourusername/AI-Digi-Ellie.git
cd AI-Digi-Ellie

2. Create a build directory:

mkdir build

3. Generate build files with CMake:

cd build; cmake ..; cd ..

4. Build the project:

cmake --build build --config Release

Windows

1. Clone the repository with submodules:

git clone --recursive https://github.com/yourusername/AI-Digi-Ellie.git
cd AI-Digi-Ellie

2. Create a build directory:

mkdir build

3. Generate build files with CMake:

cd build
cmake .. -G "Visual Studio 17 2022" -A x64
cd ..

4. Build the project:

cmake --build build --config Release

The build process will automatically download the Whisper model file (approximately 4GB) during the first build.

Prerequisites

OpenSSL Setup

1. Ensure OpenSSL 3.0 or higher is installed on your system
2. Make sure the OpenSSL development libraries are available
3. On Windows, ensure OpenSSL DLLs are in your system PATH or install OpenSSL using:

   winget install ShiningLight.OpenSSL

Opus Setup

1. Install the Opus development libraries:
   • Ubuntu/Debian: sudo apt-get install libopus-dev
   • Fedora/RHEL: sudo dnf install opus-devel
   • Arch Linux: sudo pacman -S opus
   • Windows: Opus is included with the DPP library

Ollama Setup

1. Make sure Ollama is installed and running
2. Ensure your chosen model is downloaded in Ollama (e.g., ollama pull mistral)
3. Set the DIGI_ELLIE_OLLAMA_HOST environment variable to the host of your Ollama server (default: "localhost")

Discord Setup

1. Create a Discord application and bot in the Discord Developer Portal
2. Enable Message Content Intent in the bot settings
3. Add the bot to your server with appropriate permissions (text, voice)

Azure Setup (for TTS)

1. Create an Azure account and subscribe to Azure Cognitive Services
2. Create a Speech Service resource in your preferred region
3. Get your subscription key and region from the Azure portal
4. Set the required environment variables for TTS

Running the Bot

Environment Setup

Linux/macOS

Set the required environment variables:

export DIGI_ELLIE_DISCORD_TOKEN="your_discord_token"
export DIGI_ELLIE_DEFAULT_CHANNEL_ID="your_channel_id"
export DIGI_ELLIE_AZURE_SPEECH_KEY="your_azure_speech_key"

Windows

Set the required environment variables:

set DIGI_ELLIE_DISCORD_TOKEN=your_discord_token
set DIGI_ELLIE_DEFAULT_CHANNEL_ID=your_channel_id
set DIGI_ELLIE_AZURE_SPEECH_KEY=your_azure_speech_key

Start the Whisper STT Service

The Whisper service needs to be running for voice recognition to work:

Linux/macOS

./build/bin/whisper_service

Windows

.\build\bin\Release\whisper_service.exe

You can run the Whisper service on a different machine. If you do, set the service URL in the bot:

export DIGI_ELLIE_WHISPER_SERVICE_HOST="your_whisper_server_ip"
export DIGI_ELLIE_WHISPER_SERVICE_PORT="your_whisper_server_port"

Start the bot

Linux/macOS

./build/bin/AI-Digi-Ellie

Windows

.\build\bin\Release\AI-Digi-Ellie.exe

Voice Commands

Once the bot is running:

1. Join a voice channel
2. Use the /joinvoice command to make the bot join your channel and start recording audio
3. The bot will transcribe the recorded audio using Whisper and respond to you

Project Structure

• src/ - Source files
• include/ - Header files
• vendor/ - Third-party dependencies
  • DPP/ - D++ Discord library
  • json/ - nlohmann/json library
  • spdlog/ - spdlog logging library
  • cpp-httplib/ - HTTP/HTTPS client library
  • whisper.cpp/ - Whisper speech recognition library
• whisper_models/ - Directory for Whisper model files

Troubleshooting

Windows Build Issues

• If you encounter errors about missing DLLs, ensure that OpenSSL is properly installed and in your PATH

Linux Build Issues

• Make sure all development libraries are installed (OpenSSL, Opus)
• If you encounter permission issues when running the executables, use chmod +x to make them executable

License

This project is licensed under the terms of the LICENSE file included in the repository.