Discord bot that algorithmically indexes server messages and uses Groq LLM (llama-3.1-8b-instant) to synthesise startup/project ideas
IdeaDB silently observes Discord conversations and runs a zero-LLM, zero-external-API metadata extraction pipeline over every message: keyword frequency analysis, tech-term regex matching, file attachment classification, URL platform recognition, and embed scraping. All data is stored in SQLite. When you run !ideas, it builds a structured context prompt from aggregated keywords and sends it to Groq's llama-3.1-8b-instant to generate actionable startup/project ideas with name, pitch, problem, audience, and tech stack.
Passive Discord idea-harvesting bot with algorithmic parsing and Groq LLM generation.
• Collection: listens on configured channels; parses text (stopword-filtered keyword frequency, tech-term regex), file attachments (MIME type, filename tokens, media class), URLs (known-platform table, path tokenisation, query keys), and Discord embeds.
• Storage: SQLite with guild-scoped entries table; indices on guild_id and content_type; survives restarts.
• Idea generation: Groq API (llama-3.1-8b-instant); context = top-30 keywords + content breakdown + tech terms + recent snippets; returns structured ideas with name, pitch, problem, audience, tech stack.
• Commands: !ideas [count], !keywords, !stats, !clear (admin).
• Auto-timer: generates 3 ideas every N minutes (configurable) only if new messages have arrived since last run.
• Stack: Python 3.11+, discord.py ≥ 2.3, groq-python, aiohttp, python-dotenv, SQLite (stdlib).