Discord bot ships with a rogue’s gallery of AI personas

What it does

This Python bot connects a Discord server to the Groq API for fast LLM responses. It watches for mentions, direct messages, and replies in channels toggled active via /toggleactive, and can be configured to speak one of twelve interface languages.

The interesting bit

Instead of a single system prompt, the bot offers seven canned personas—DAN, Ivan, Luna, Suzume, and others—each with a distinct voice and moral compass, selectable through config.yml. You can also write your own instructions in a plain .txt file, making this essentially glue between discord.py and Groq with a personality vending machine bolted on top.

Key highlights

• Hybrid slash and normal command support
• Mention recognition and channel-specific /toggleactive mode
• Seven built-in personalities plus custom instruction files
• Twelve interface languages from Turkish to Polish
• Environment-variable credential management and Docker support

Caveats

• The authors admit they are “too lazy to list all” commands beyond /help
• Documentation mixes config.yml and config.json references with stale line numbers
• Several built-in personas are explicitly unfiltered or explicit, which may surprise server moderators

Verdict

Worth a look for Discord admins who want a drop-in LLM companion with personality presets. Avoid if you need rigorous docs, command transparency, or strict content moderation.