Discord OpenAI Bot

A lightweight Discord bot powered by OpenAI’s Chat Completions API. It exposes a /ask slash command that creates a dedicated thread for each request, keeps the entire thread history in memory, and replies inside that thread so everyone can follow the conversation.

Status: Actively developed proof-of-concept. Great for personal servers; adapt before running at larger scale or with persistent storage.

Features

• /ask slash command with per-thread OpenAI context
• Automatic thread creation and cleanup when threads are archived/deleted
• In-memory conversation history for quick experimentation
• Configurable OpenAI model via OPENAI_MODEL

Prerequisites

• Python 3.11+
• Discord application & bot token with the Message Content Intent enabled
• OpenAI API key with access to the model you configure (default gpt-4o-mini)

Quick Start

1. Clone and install deps:

   git clone https://github.com/noahatkins/python-bot.git
   cd python-bot
   python -m venv .venv
   .\.venv\Scripts\activate  # source .venv/bin/activate on macOS/Linux
   pip install -r requirements.txt

2. Copy env.example → .env and fill in:
   • DISCORD_TOKEN
   • OPENAI_API_KEY
   • optional OPENAI_MODEL
3. Run the bot:

   python bot.py

4. Invite the bot with scopes bot applications.commands. Grant Send Messages, Create Public Threads, Send Messages in Threads, Manage Threads, Read Message History, and Use Slash Commands.

/ask Workflow

1. /ask prompt:"Could you outline the steps to deploy this Discord bot to a Raspberry Pi?"
2. Bot replies in the channel, creates a thread, asks OpenAI, and posts the first answer inside the new thread.
3. Keep chatting in that thread; every new user message becomes part of the OpenAI context so responses stay on topic.

The bot ignores messages outside those threads and skips replies to other bots by default.

Configuration

Variable	Description	Required
DISCORD_TOKEN	Token from the Discord developer portal	✅
OPENAI_API_KEY	OpenAI API key with chat-completions access	✅
OPENAI_MODEL	Override default model (gpt-4o-mini)	❌

Secrets are loaded via .env thanks to python-dotenv. Never commit .env; use env.example when sharing config.

Secret Hygiene

• .env is already gitignored; keep your real keys there.
• Rotate Discord/OpenAI keys immediately if they leak (Discord dev portal → Regenerate Token, OpenAI dashboard → View API keys).
• When opening PRs, redact logs that might contain prompt contents or server names.

Development

• Formatting/linting: python -m py_compile bot.py services/openai_client.py
• Run locally with python bot.py
• Tests are manual for now; add your own harness or mocks before shipping to production servers.

Troubleshooting

• Slash command not visible: restart the bot so bot.tree.sync() runs, and ensure the invite included applications.commands.
• Thread creation fails: verify the bot has the thread permissions listed above and that you run /ask in a guild text channel (not DMs).
• OpenAI errors: check that OPENAI_API_KEY is valid and the model name exists in your account. The bot logs failures to the console.

License

MIT © noahatkins