This project is a Discord bot that uses AI models through Ollama API to generate contextual responses. It maintains both conversation history with the bot and recent channel message history to understand conversation context even when joining discussions mid-conversation.
Caution
This project uses a selfbot which violates Discord's Terms of Service (ToS). I am NOT responsible if your Discord account gets suspended or banned for using this code. Always use a secondary/alternative account token for this project to avoid losing access to your main account.
Before you begin, ensure you have the following installed:
- Node.js (v20.18 or higher)
(v22.12.0 used for dev) - Ollama - Local AI model runner
- A Discord account and your Discord token (How to get your Discord token)
-
Clone the repository:
git clone https://github.com/MatisAgr/Easy-SelfBot-Discord-IA.git -
Navigate to the project folder:
cd Easy-SelfBot-Discord-IA -
Install dependencies:
npm i
- Rename the
.env.examplefile to.env. - Fill in the environment variables in the
.envfile:DISCORD_TOKEN: Your Discord token (get it here).HISTORY_LIMIT: The number of bot conversation pairs to keep in memory (default: 10).USE_HISTORY_CONVERSATION: Whether to enable channel message history tracking (default: true).NB_MESSAGES_HISTORY: Number of recent channel messages to keep for context (default: 10).INACTIVITY_TIMEOUT_MINUTES: Minutes before resetting channel memory due to inactivity (default: 0).OLLAMA_API_URL: The URL of the Ollama API (default: "http://localhost:11434").OLLAMA_MODEL: The name of the model to use with Ollama (check available models withollama list).BOT_CONTEXT: The context for the bot when generating responses.SHOW_THINKING: Whether to show the bot's thinking process in Discord messages (default: false).LOG_CONVERSATIONS: Whether to log conversations to a file (default: false).LOGS_FILE_PATH: Path to the conversation logs file (default: "./conversations_log.json").DEBUG_MODE: Whether to save the bot's memory (prompts) to files for debugging (default: false).MEMORY_FOLDER: Directory where memory files will be stored (default: "./debug_bot_memory").
When DEBUG_MODE is enabled, the bot will save its raw prompts in the specified MEMORY_FOLDER organized by:
- Server name
- Channel name
- Timestamp and author
This feature helps you debug how the bot's "memory" works and what data is being sent to the AI model. Files are organized with the following structure:
debug_bot_memory/
├── ServerName/
│ ├── channel-name/
│ │ ├── raw_prompt_username_123456.txt
│ │ └── raw_prompt_username_123457.txt
└── DirectMessages/
└── direct/
└── raw_prompt_username_123458.txt
Each file contains the exact prompt sent to Ollama, including the context, conversation history, and current message.
To start the bot, run the following command:
node start.js
The bot will connect to your Discord account and start listening for messages. It will respond to messages where it's mentioned and automatically track channel conversations for context.
Key behaviors:
- Responds only when explicitly mentioned (@) or replied to
- Tracks all channel messages in real-time for context (optional)
- Maintains separate memory per channel
- Automatically manages memory limits to prevent overflow
- Resets inactive channel memories after 0 minutes (configurable)
This project uses the following dependencies:
discord.js-selfbot-v13: To interact with the Discord API.dotenv: To manage environment variables.axios: To make HTTP requests to the Ollama API.
- Node.js - JavaScript runtime
- Discord.js-selfbot-v13 - Discord API wrapper for user accounts
- Ollama - Open-source model runner for LLMs
- Environment Variables - For secure configuration management