TeaL;DR

TeaL;DR makes Discord searchable with AI. Instead of keyword searching, users ask natural questions and get instant answers about what happened in their communities.

Comment

TeaL;DR - AI-Powered Discord Knowledge Bot

Inspiration

The inspiration for TeaL;DR came from a fundamental pain point in Discord communities: searching is limited to keywords and keeping up with active open-source communities is time-consuming. While Discord is the hub for countless projects and discussions, finding specific information or understanding what happened while you were away requires manually scrolling through endless messages. TeaL;DR solves this by creating an intelligent knowledge base that understands natural language queries and provides instant, contextual answers.

What It Does

TeaL;DR transforms Discord servers into searchable knowledge bases by:

  • Semantic Search: Uses AI embeddings to understand the meaning behind messages, not just keywords
  • Natural Language Queries: Ask questions like "What did the team discuss about the API update?" instead of searching for specific terms
  • Smart Recaps: Automatically generates summaries of server activity over custom timeframes
  • Multi-Server Support: Works across multiple Discord servers with intelligent server selection
  • DM Commands: Allows users to query servers directly from DMs with interactive server picker
  • Real-time Updates: Continuously indexes new messages to keep the knowledge base current

How We Built It

  • Backend: Python with Discord.py for bot functionality
  • AI/ML: Google Gemini API for intelligent responses and semantic understanding
  • Database: Supabase for storing message embeddings and metadata
  • Vector Search: Custom embedding system for semantic message retrieval
  • Infrastructure: Deployed on Google Cloud Platform's free tier (f1-micro Compute Engine)
  • Frontend: Modern web interface for

Challenges We Ran Into

  • Storage Optimization: Supabase storage costs for message embeddings in large servers
  • Scalability: Managing embeddings for thousands of messages while maintaining fast query performance
  • Discord API Limits: Working within rate limits and message retrieval constraints
  • Deployment: Moving from paid services (Render/Railway) to completely free infrastructure
  • Multi-Server Coordination: Handling simultaneous requests across multiple Discord servers

Accomplishments We're Proud Of

  • Community Adoption: Users are actively inviting TeaL;DR to their servers, showing real demand
  • Free Infrastructure: Successfully deployed on GCP's free tier, making it sustainable for long-term growth
  • Advanced Features: Implemented DM commands, multi-server search, and interactive server selection
  • Performance: Optimized embedding storage and retrieval for fast responses
  • User First Approach: Created intuitive commands and comprehensive documentation

What We Learned

  • Free Tier Engineering: Learned to optimize for resource-constrained environments (f1-micro with 0.6GB RAM)
  • Vector Database Design: Understanding how to efficiently store and query embeddings at scale
  • Discord Bot Architecture: Best practices for building robust, scalable Discord applications
  • Cost Optimization: Techniques for minimizing infrastructure costs while maintaining performance
  • User-Centric Design: Importance of intuitive commands and comprehensive documentation

What's Next for TeaL;DR

  • Voice Transcriptions: Add support for voice channel transcription and search
  • Advanced Analytics: Dashboard for server activity insights and bot usage statistics
  • Multi-Modal Search: Support for images, files, and other media types
  • Community Features: Server-to-server knowledge sharing and collaborative filtering
  • Enterprise Features: Team management, custom models, and advanced security features
  • Mobile App: Native mobile application for on-the-go server monitoring and querying

Built With

Share this project:

Updates