LumiaBot

A Discord bot built with Bun + TypeScript that provides AI-powered chat completions with optional web search capabilities via SearXNG.

Features

AI Chat Completions: Powered by OpenAI's GPT models
Web Search Integration: Uses SearXNG for real-time web search results
Message Triggers: Responds to mentions (@bot) and keywords ("Bad Kitty", "Lumia")
Slash Commands: Modern Discord slash command interface
Streaming Support: Real-time response streaming (optional)
TypeScript: Fully typed for better development experience
Bun Runtime: Fast, modern JavaScript runtime

Prerequisites

Bun installed (v1.3.5 or higher)
A Discord Application with Bot token
OpenAI API key
SearXNG instance (public or self-hosted)

Installation

Clone the repository (or create from scratch):

cd LumiaBot

Install dependencies:

bun install

Configure environment variables:

cp .env.example .env
# Edit .env with your actual credentials

Set up bot definitions (IMPORTANT):

# Create the prompt_storage directory structure
mkdir -p prompt_storage/persona prompt_storage/instructions prompt_storage/config

# Copy example templates from prompt_storage.example/
cp -r prompt_storage.example/* prompt_storage/

# Edit the files to create your own bot personality
# See BOT_SETUP.md for detailed instructions

Configuration

Create a .env file with the following variables:

# Discord Bot Configuration
DISCORD_TOKEN=your_discord_bot_token_here
DISCORD_CLIENT_ID=your_discord_application_id_here
DISCORD_CLIENT_SECRET=your_discord_client_secret_here
DISCORD_REDIRECT_URI=http://localhost:3000/auth/callback

# Bot Identity Configuration (Optional)
# These values are used as template variables in prompt_storage files
# BOT_NAME=Bad Kitty                    # Bot's display name
# BOT_OWNER_NAME=Prolix                 # Owner's name
# BOT_OWNER_ID=944783522059673691       # Owner's Discord ID
# BOT_OWNER_USERNAME=prolix_oc          # Owner's Discord username

# OpenAI Configuration
OPENAI_API_KEY=your_openai_api_key_here
# OPENAI_BASE_URL=https://api.openai.com/v1  # Optional: Custom API base URL (e.g., for OpenRouter, Together AI)
OPENAI_MODEL=gpt-4o-mini
# OPENAI_MODEL_ALIAS=gpt-4o-mini  # Optional: Map to a different model name for the provider
OPENAI_MAX_TOKENS=2000
OPENAI_TEMPERATURE=0.7
OPENAI_FILTER_REASONING=true  # Filter out reasoning content from responses (e.g., o1/o3 models)

# SearXNG Configuration
SEARXNG_URL=https://search.example.com
SEARXNG_MAX_RESULTS=5
SEARXNG_SAFE_SEARCH=1

# Server Configuration
PORT=3000

Bot Identity Configuration

You can customize your bot's identity and owner information through environment variables. These values are used as template variables (e.g., {botName}, {ownerName}) in your prompt_storage files:

# Optional - Bot identity (defaults shown)
BOT_NAME=Bad Kitty                    # Bot's display name
BOT_OWNER_NAME=Prolix                 # Owner's name
BOT_OWNER_ID=944783522059673691       # Owner's Discord ID
BOT_OWNER_USERNAME=prolix_oc          # Owner's Discord username

These variables allow you to:

Change the bot's name without editing prompt files
Set the bot's owner for special recognition in prompts
Maintain consistent identity across all prompt templates

See BOT_SETUP.md for more details on using template variables in your bot's personality files.

Using Custom AI Providers

The bot supports custom OpenAI-compatible API providers such as:

OpenRouter (access to multiple models)
Together AI
Groq
Google GenAI (supports video/audio modality)
Local models (via llama.cpp, etc.)

Example: OpenRouter Configuration

OPENAI_BASE_URL=https://openrouter.ai/api/v1
OPENAI_API_KEY=your_openrouter_key_here
OPENAI_MODEL=anthropic/claude-sonnet-4-5
OPENAI_MODEL_ALIAS=claude-sonnet-4-5  # Optional: alias for the model

Example: Local Model via llama.cpp

OPENAI_BASE_URL=http://localhost:8080/v1
OPENAI_API_KEY=optional-for-local
OPENAI_MODEL=local-model

Using Google GenAI (Direct Gemini API)

For Gemini 3 models, you can use Google's native GenAI SDK instead of going through an OpenAI-compatible proxy. This provides better native support for Gemini-specific features.

When to use this:

When using Gemini 3 Flash/Pro models
When you have direct access to Google's Gemini API
When you want to avoid proxy layers

Configuration:

# Set the model to a Gemini 3 variant
OPENAI_MODEL=gemini-3-flash

# Configure Google GenAI (required to activate the native SDK)
GEMINI_API_KEY=your_gemini_api_key_here  # Get from https://ai.google.dev/
# GEMINI_BASE_URL=https://generativelanguage.googleapis.com  # Optional: for custom endpoints

How it works:

If GEMINI_API_KEY is set AND the model name contains "gemini-3", the bot automatically uses the Google GenAI SDK
Otherwise, it falls back to the OpenAI SDK (for OpenRouter, Together AI, etc.)
The bot logs which service it's using on startup: [AI Service] Using Google GenAI for gemini-3-flash

Getting Discord Credentials

Go to Discord Developer Portal
Create a new application
Go to "Bot" section and create a bot
Copy the bot token (DISCORD_TOKEN)
Go to "OAuth2" → "General" and copy the Client ID (DISCORD_CLIENT_ID) and Client Secret (DISCORD_CLIENT_SECRET)
In "OAuth2" → "URL Generator", select bot and applications.commands scopes, then copy the generated URL to invite your bot

Getting OpenAI API Key

Go to OpenAI Platform
Create an account or sign in
Go to API keys section and create a new secret key

Setting up SearXNG

You have two options:

Option 1: Use a public instance

Find public instances at searx.space
Note: Public instances may have rate limits

Option 2: Self-host SearXNG

docker run -d --name searxng -p 8080:8080 -e "BASE_URL=http://localhost:8080" searxng/searxng

Usage

Register Slash Commands

Before using the bot, register the slash commands with Discord:

bun run register-commands

Start the Bot

Development mode (with hot reload):

bun run dev

Production mode:

bun run start

Available Commands

Once the bot is running and invited to your server, use these slash commands:

/chat <message> [search] - Chat with the AI
- message: Your question or message
- search: (Optional) Enable web search for better answers
/search <query> [category] [timerange] - Search the web
- query: What to search for
- category: (Optional) Filter by category (general, images, news, science, files)
- timerange: (Optional) Filter by time (day, month, year)

Project Structure

LumiaBot/
├── src/
│   ├── bot/
│   │   └── client.ts          # Discord client setup
│   ├── commands/
│   │   ├── chat.ts            # /chat command
│   │   └── search.ts          # /search command
│   ├── scripts/
│   │   └── register-commands.ts # Command registration script
│   ├── services/
│   │   ├── openai.ts          # OpenAI integration
│   │   └── searxng.ts         # SearXNG search integration
│   ├── utils/
│   │   └── config.ts          # Configuration management
│   └── index.ts               # Entry point
├── .env                       # Environment variables
├── .env.example               # Example environment file
├── package.json
├── tsconfig.json
└── README.md

Development

Adding New Commands

Create a new file in src/commands/:

import { SlashCommandBuilder, ChatInputCommandInteraction } from 'discord.js';
import type { Command } from '../bot/client';

const command: Command = {
  data: new SlashCommandBuilder()
    .setName('mycommand')
    .setDescription('Description of my command'),
  
  async execute(interaction: ChatInputCommandInteraction) {
    await interaction.reply('Hello from my command!');
  },
};

export default command;

Register the command:

bun run register-commands

Environment Variables

All configuration is managed through environment variables. See .env.example for all available options.

Bot Personality Configuration

The bot's personality is defined in the prompt_storage/ directory, which is excluded from git for privacy. You must create these files yourself using the provided templates.

Setting Up Bot Definitions

Create the directory structure:

mkdir -p prompt_storage/persona prompt_storage/instructions prompt_storage/config

Copy example templates:

cp -r prompt_storage.example/* prompt_storage/

Customize the files to create your own bot personality (see BOT_SETUP.md for detailed instructions)
Restart the bot to load your custom personality

Why This Approach?

The prompt_storage/ directory is gitignored to keep your bot's unique personality private. The example templates in prompt_storage.example/ show you the required file structure and format without exposing any proprietary content.

File Structure

prompt_storage/
├── persona/
│   ├── identity.txt           # Main bot personality/system prompt
│   ├── boredom_pings.json     # Random messages when bot is "bored"
│   ├── error_templates.json   # Error/fallback responses
│   └── command_responses.json # Command-specific responses
├── instructions/
│   ├── video_reaction.txt     # How to react to videos
│   ├── reply_context.json     # Reply conversation templates
│   ├── memory_system.txt      # Memory formation instructions
│   └── boredom_updates.txt    # Boredom opt-in/opt-out responses
└── config/
    ├── triggers.json          # Bot trigger keywords
    └── tool_descriptions.json # Tool descriptions with personality

See BOT_SETUP.md for complete documentation on creating and customizing these files.

Message Triggers

The bot automatically responds to messages when:

Mentioned directly - @BadKittyBot hello!
Trigger keywords detected - Messages containing:
- Bad Kitty (case insensitive)
- Lumia (case insensitive)
Replied to — replies to their messages

Examples

@BadKittyBot What's the weather today?
→ Bot responds with AI-generated answer

Hey Bad Kitty, can you help me with something?
→ Bot responds with AI-generated answer

Lumia, what do you think about this?
→ Bot responds with AI-generated answer

How it Works

The bot monitors all messages in channels it has access to
When triggered, it extracts the actual message content (removing the trigger)
Generates an AI response using the bot's personality from prompt_storage/persona/identity.txt
Replies directly to the user's message

Note: Message triggers do not use web search by default (unlike the /chat command with search: true).

Reasoning Content Filtering

The bot automatically filters out reasoning content from AI models that include it (such as o1, o3, DeepSeek-R1, etc.). This prevents internal "thinking" from being shown to users.

Filtered patterns include:

<think>...</think> tags
<reasoning>...</reasoning> tags
[REASONING]...[/REASONING] tags
Lines starting with "reasoning:", "thinking:", etc.

To disable filtering, set:

OPENAI_FILTER_REASONING=false

Troubleshooting

Bot not responding to commands

Ensure you've registered commands with bun run register-commands
Check that the bot has proper permissions in the Discord server
Verify the bot token is correct

OpenAI errors

Check your API key is valid and has available credits
Verify the model name is correct
Check rate limits haven't been exceeded

SearXNG errors

Ensure the SearXNG URL is accessible
Check if the instance requires authentication
Verify the instance supports JSON output format

License

MIT

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.

Name		Name	Last commit message	Last commit date
Latest commit History 17 Commits
knowledge_documents		knowledge_documents
prompt_storage.example		prompt_storage.example
src		src
.env.example		.env.example
.gitignore		.gitignore
BOT_SETUP.md		BOT_SETUP.md
LICENSE.md		LICENSE.md
README.md		README.md
bun.lock		bun.lock
index.ts		index.ts
package.json		package.json
tsconfig.json		tsconfig.json

Folders and files

Latest commit

History

Repository files navigation