How adnanquazi/zap works
Project type and purpose
• This is a full-stack, microservices-style application built to integrate with Slack.
• Purpose: Provide a “knowledge assistant” (Zap) in Slack channels that answers questions using channel history, uploaded documents, and an embedding-backed LLM.
File structure
• Top‐level monorepo managed by Turbo/pnpm workspaces.
• apps/client: Next.js app hosting the marketing site & Slack install/oauth pages.
• apps/server: Node.js app handling Slack OAuth, event subscriptions, slash commands, business logic, rate‐limiting, context pipelines.
• services/embedding-service: Python FastAPI (or similar) microservice that computes embeddings.
• supabase: Database schema & migrations for storing messages, user settings.
• Redis used for rate limits, context cooldown, concurrency limiting.
README for additional insights
• Built with Next.js, Node.js, Supabase, Redis, Python embedding service, Modal (for serverless?).
• Key flows: OAuth, installing Slack app, event webhook processing, slash‐command handlers, privacy controls, embedding pipeline.