About

Work-In-Progress discord bot written with Discord.js.

Capabilities

Slash commands:

• /mc - Returns information about the specified Minecraft server (uses mcsrvstat)
• /ping - Returns latency stats
• /ai - Generate an AI response without mentioning the bot, optionally without a prompt for use of bare Gemini
• /manage_ai - Change the AI prompt during runtime, dump conversation history, wipe conversation history
• /whatis - Returns the definition of a specified word (uses Urban Dictionary)

AI

Will respond to you like an AI chat-bot if you mention the bot user, or with the /ai command.

Setup packages

First, ensure you have all the production dependencies:
(I use pnpm, you could substitute it for npm)

pnpm install --production

You may encounter errors with installing the optional dependencies. This is okay. They are not required, just used to improve performance.
Learn more about these dependencies here.

Configure

Create a file called .env in the project directory. Here is the outline of the file. Replace any <> fields with their respective values.

TOKEN=<bot_token>
CLIENT_ID=<bot_client_id>
GUILD_ID=<dev_guild_id> # OPTIONAL: When not deploying commands globally, deploys commands to this guild for development purposes.
RAPIDAPI_KEY=<key> # Needed to use the /whatis command. Get a key here: https://rapidapi.com/community/api/urban-dictionary
GOOGLEAI_KEY=<key> # Needed to use AI functionality. Get a key here: https://ai.google.dev/tutorials/setup
GOOGLEAI_PROMPT=<prompt> # Tells the AI what to do and how to act. Example: "You are a fish. You speak only like a fish, you speak no other way."
AI_HUMANS_ONLY=<trueorfalse> # Whether the AI responds when other bots/apps mention it

Deploy slash commands

You will need to deploy slash commands for them to be available for use:

pnpm run deploy-commands true

This deploys the commands globally; to every server the bot has joined. If you wish to deploy to a specified server (GUILD_ID in .env), for say, development purposes, run:

pnpm run deploy-commands false

Running

To run the bot, run:

node .

You can also run it with Bun. As of Bun 1.1.5, everything works except for TextDecoderStream used in the Generative AI Package and the Brotli implementation used in Axios. Currently, i've implemented a workaround or Axios at run-time. To make TextDecoderStream in Google's package work, see this issue.