Discord-Indexer

A Discord bot that indexes server message history and provides intelligent search through direct messages using local LLM processing.

What it does

The bot extracts all messages from Discord channels, converts them to searchable vectors, and allows users to query the conversation history through natural language. Instead of manually scrolling through months of chat history, users can ask questions like "What did we decide about the deployment schedule?" and get relevant message context with source links.

Core Features

• Message Indexing: Automatically processes Discord channel history with metadata preservation
• Semantic Search: Vector-based search using sentence transformers for contextual relevance
• LLM Integration: Local Mistral 7B model for response generation and context synthesis
• DM Interface: Query the indexed content through direct messages with the bot
• Voice Processing: Speech-to-text and text-to-speech support via Chatterbox integration
• Real-time Processing: Continuous indexing of new messages as they arrive

Technical Architecture

Backend Stack

• Discord.py: Bot framework and API integration
• FastAPI: REST API for message processing pipeline
• ChromaDB: Vector database for message storage and retrieval
• Sentence Transformers: Text embedding generation
• Mistral 7B: Local language model inference

Data Pipeline

1. Discord message extraction with pagination handling
2. Text preprocessing and metadata extraction
3. Vector embedding generation and storage
4. RAG (Retrieval-Augmented Generation) for query processing
5. Response formatting for Discord's message constraints

Hardware Requirements

• GPU: RTX 3090 or equivalent (tested configuration)
• RAM: 16GB+ recommended for local LLM inference
• Storage: Varies based on server message volume

Development Timeline

The project follows a 10-week sprint plan focusing on:

• Weeks 1-4: Core message extraction and vector storage
• Weeks 5-6: LLM integration and RAG implementation
• Weeks 7-8: Discord interface and query processing
• Weeks 9-10: Performance optimization and voice features

Current Status

This project is in active development. Core functionality targets include reliable indexing of 50+ messages, sub-5-second query response times, and stable operation on consumer hardware.

Installation & Setup

Prerequisites

1. Download & Install Ollama --> (https://ollama.com/download/windows)

2. Open up a terminal and run this command:

   ollama pull llama3.1:8b

3. Install Python dependencies from the root directory:

   pip install -r requirements.txt

⚠️ Important for Existing Developers

If you have previously run the bot, you MUST delete your existing database before proceeding:

# Delete the entire databases directory
rm -rf src/db/databases
# Or on Windows:
rmdir /s src\db\databases

This is required due to configuration system changes in the current development phase.

Setup Instructions

1. Configure your Discord bot by running the setup script:

   python setup_bot.py

   This will guide you through:

   • Entering your Discord server ID and name
   • Configuring message processing error handling
   • Setting up per-server configuration

2. Start the bot:

   python main.py

   The bot will now only process messages from configured servers.

Configuration Management

After initial setup, you can modify bot settings using:

# Modify settings for a specific server
python change_settings.py [server_id]

# Or run without server_id to select from available servers
python change_settings.py

Development Notes

• The bot requires server configuration before it will start
• Only configured servers will have their messages indexed
• Each Discord server maintains separate configuration and database storage
• Configuration is stored persistently in ChromaDB

Usage

Usage documentation will be provided upon initial release.

Contributing

Contribution guidelines will be established as the project structure stabilizes.