Local LLM Discord Chatbot Powered by Open WebUI

This is a modular and customizable Discord chatbot that connects to your local LLM using Open WebUI via REST API.

It supports conversational memory, system personas, summarization, logging, and stat tracking — all configurable through a .env file.

---

Features

• !ask — Send a prompt to your local LLM and get a response
• !summarize — Send text to a separate summarization model
• !setpersona — Define how the assistant should behave for your user session
• !reset — Clear your session memory
• !stats — View global usage stats (token count, requests, latency)
• !debug — Toggle debug mode (enables file logging)
• !commands — List all available commands

---

How It Works

This bot connects to Open WebUI using its REST API:

• Each command (!ask, !summarize, etc.) sends messages to the /api/chat/completions endpoint
• You can specify different model values in the request payload (e.g. "discord", "summarizer")
• User history and personas are tracked in memory per Discord user

---

Requirements

• Python 3.8+
• A local or network-accessible instance of Open WebUI
• A Discord bot token

---

Setup

1. Clone the repo
2. Install dependencies:

pip install -r requirements.txt

3. Create a .env file in the root directory:

DISCORD_BOT_TOKEN=your_discord_bot_token
OPEN_WEBUI_API_TOKEN=your_open_webui_token
OPEN_WEBUI_API_URL=http://localhost:3000/api/chat/completions
DEBUG_MODE=True

4. Run the bot:

python bot.py

---

Notes

• Debug mode logs all requests/responses and usage metrics to botlog.log
• User history is stored in-memory (non-persistent)
• You can customize model names, persona behavior, and command cooldowns easily in the code

---

License

MIT — use freely, but attribution is appreciated.

---

Contributions Welcome