Ollama Bot

Table of Contents

• Ollama Bot
• Table of Contents
• Introduction
• Usage
  • Commands
  • Bot Interaction
• Installation
• File Structure
• Local Development
  • Running the Bot
  • Running with Docker

Introduction

A Discord bot that integrates with Ollama to generate AI-powered replies.

This bot allows you to interact with Ollama's models directly from Discord, making it easy to generate text responses, summaries, and more.
Both the Discord bot and Ollama are designed to be run locally in Docker containers, providing a self-hosted solution for AI interactions, ensuring privacy and control over your data.

Usage

Commands

Command	Arguments	Description
/settings model set <model>	model (string)	Set the Ollama model for this guild.
/settings model get	none	Get the current Ollama model.
/settings model list	none	List all available models from the Ollama server.
/settings model pull <model>	model (string)	Download (pull) a model into Ollama.
/settings prompt set <prompt>	prompt (string)	Set the system prompt sent at the start of each chat.
/settings prompt get	none	Get the current system prompt.
/settings role set [role]	role (@Role) or leave blank	Restrict commands to users who have this role (or clear).
/settings role get	none	Show the currently required role (if any).
/settings history set [reply_history] [message_history]	reply_history (int), message_history (int)	Set how many messages to fetch as context.
/settings history get	none	Show current reply & message history limits.

Bot Interaction

To chat with the bot you can do one of the following:

• Mention the bot: Mention the bot (example: @Ollama Bot) followed by your message.
• Reply to the bot: Click on the bot's message and type your reply.

Doing either will send your message to the bot, which will then generate a response using the configured Ollama model.

Installation

To install, run, and add the bot to your Discord server, follow these steps:

1. Create a Discord Application:
   • Go to the Discord Developer Portal.
   • In the Bot section:
     • Note down the Token, as you will need it to run the bot.
     • Under Privileged Gateway Intents, enable the Message Content Intent to allow the bot to read messages.
2. Invite the Bot to Your Server:
   • In the Installation section:
     • Under Installation Contexts, Select Guild Install only.
     • Under Default Install Settings, add the applications.commands and bot scopes, then the Send Messages and Send Messages in Threads permissions.
     • Finally, copy the Discord Provided Link and paste it into your browser to invite the bot to your server.

Tip

For more information on creating a Discord bot, refer to the Discord Developer Portal documentation.

3. Run Docker Compose
   • Ensure you have Docker installed.
   • Copy the docker-compose.yml file from this repository to your local machine.
   • Customize the docker-compose.yml file if necessary, such as changing mounted volumes or ports.
   • Create a .env file in the root directory with your Discord bot token and Ollama API URL:

     DISCORD_APP_TOKEN=your_discord_app_token
     OLLAMA_API_URL=your_ollama_api_url

   • Run the following command to start the bot and Ollama (detached):

     docker compose up -d

Tip

For more information about the Ollama Docker image, refer to the Ollama documentation (GPU support is available here).

File Structure

.
├── .env                     The environment variables for the bot
├── .gitignore               Files and directories to ignore in git
├── README.md                Documentation for the bot
├── Dockerfile               Dockerfile to build the bot image
├── docker-compose.yml       Example Docker Compose file
├── src/                     Source code for the bot
│   ├── main.py              Main entry point for the bot
│   ├── ollama_bot.py        Discord client and message handler
│   ├── settings_manager.py  Manages settings for the bot
│   ├── history_group.py     /settings history subcommands
│   ├── model_group.py       /settings model subcommands
│   ├── prompt_group.py      /settings prompt subcommands
│   ├── role_group.py        /settings role subcommands
│   ├── settings_group.py    /settings group command
│   ├── utils.py             Utility functions for the bot
│   └── requirements.txt     Python dependencies

Local Development

Running the Bot

Important

Ollama must be running for the bot to function properly. You can run Ollama using Docker with the following command:

docker run -d -v ollama:/root/.ollama -p 11434:11434 --name ollama ollama/ollama

To run the bot locally, follow these steps:

1. Clone the repository:

   git clone https://github.com/TechnoBro03/OllamaBot
   cd ollamabot

2. Create & activate a virtual environment (optional but recommended):

   python -m venv venv
   source venv/bin/activate # On Windows use `venv\Scripts\activate`

3. Install dependencies:

   pip install -r src/requirements.txt

4. Create a .env file in the root directory with the following content:

   DISCORD_APP_TOKEN=your_discord_app_token
   OLLAMA_API_URL=your_ollama_api_url

5. Run the bot:

   python src/main.py

Running with Docker

To run the bot using Docker, follow these steps:

1. Ensure you have Docker installed on your machine.
2. Create a .env file in the root directory with your Discord bot token and Ollama API URL:

   DISCORD_APP_TOKEN=your_discord_app_token
   OLLAMA_API_URL=your_ollama_api_url

3. Build and run the Docker container:

   docker build -t ollama-bot .
   docker run -d --env-file .env --name ollama-bot ollama-bot