A Discord bot that captures, analyzes, and provides intelligent insights from Discord server conversations using AI-powered summarization and semantic search.

• Real-time capture: Monitor all messages across channels with metadata (timestamp, user, channel)
• Historical backfill: Import existing server message history
• Organized storage: PostgreSQL database with efficient indexing by channel and user
• Metadata preservation: Track message edits, reactions, attachments, and thread context

• Smart summarization: Generate channel summaries for any time period
• Semantic search: Ask natural language questions about server activity
• Vector embeddings: Store message embeddings for similarity-based retrieval
• Context-aware responses: LLM analyzes relevant messages to answer queries

• Prefix commands: Easy-to-use Discord commands for summaries and searches
• Channel filtering: Focus analysis on specific channels or users
• Time-based queries: Filter by date ranges (last 24h, week, month, custom)
• Export capabilities: Generate reports and data exports

• Language: Python 3.11+ with asyncio
• Discord Library: discord.py 2.x with proper intents
• Database: PostgreSQL for relational data (messages, users, channels)
• ORM: SQLAlchemy 2.0+ with asyncio support for database operations
• Vector Database: ChromaDB for embeddings and similarity search
• AI Framework: LangChain for RAG operations and agent functionality
• AI Services: OpenAI API (GPT-4 + text-embedding-3-small)
• Package Manager: uv (as per user preference)

-- PostgreSQL: Core tables for message organization
users (id, discord_id, username, display_name, joined_at)
channels (id, discord_id, name, guild_id, type)
messages (id, discord_id, user_id, channel_id, content, timestamp)
channel_summaries (id, channel_id, period, summary, generated_at)

-- ChromaDB: Vector embeddings with metadata
-- Collections organized by guild/channel for efficient retrieval
-- Metadata includes: message_id, user_id, channel_id, timestamp

1. Message Listener: Captures real-time messages via Discord events
2. Storage Engine: Async database operations with proper indexing
3. Embedding Service: Batch processing for vector embeddings
4. Query Engine: Semantic search and LLM integration
5. Command Interface: Prefix commands for user interaction

• Set up development environment with uv
• Configure PostgreSQL for relational data
• Set up ChromaDB for vector storage
• Basic Discord bot with message capture
• Database schema and migrations

• Message storage pipeline
• Historical message backfill
• Basic filtering and organization
• Simple query commands

• OpenAI API integration
• Embedding generation and storage
• Semantic search implementation
• Summarization features

• Command refinement
• Error handling and logging
• Security and privacy features
• Deployment setup

# Get message history
!history @username limit:50 since:24h
!history #general since:7d

# Get AI-powered summaries
!summarize @username limit:100 since:7d
!summarize #dev-chat since:24h
!summarize limit:50 since:1h

# Commands support flexible filtering:
# - User mentions: @username
# - Channel mentions: #channel-name
# - Time ranges: since:1h, since:7d, since:30m
# - Result limits: limit:10, limit:50, limit:100

• Optional opt-out mechanism for users/channels
• Encrypted token storage
• GDPR-compliant data handling
• Permission-based access control

• Accurate message capture (99%+ reliability)
• Fast query responses (<3 seconds)
• Meaningful summaries and insights
• Easy-to-use interface for server admins

The project was successfully migrated from direct asyncpg usage to SQLAlchemy 2.0 with async support for improved maintainability and type safety.

Key Benefits:

• ORM Models: Type-safe database models with relationships
• Query Builder: Declarative query construction with compile-time validation
• Connection Management: Built-in connection pooling and session management
• Migrations: Future-ready for Alembic migrations
• Async Support: Full asyncio compatibility with async/await patterns

Database Models:

• Server: Discord guild/server information
• Channel: Channel metadata with server relationships
• User: User profiles with discriminator support
• ServerMember: Many-to-many server membership with roles (JSONB)
• Message: Full message content with attachments, embeds, reactions (JSONB)

Usage with uv:

# Install dependencies
uv sync

# Initialize database (requires PostgreSQL connection)
uv run python init_db.py

# Start the Discord bot
uv run python bot.py

Environment Variables Required:

• DB_HOST: PostgreSQL host
• DB_PORT: PostgreSQL port (default: 5432)
• DB_NAME: Database name
• DB_USER: Database username
• DB_PASSWORD: Database password
• DISCORD_TOKEN: Discord bot token