Lumina - Enterprise AI Knowledge Hub

A high-fidelity RAG platform for enterprise document search and chat.

Tech Stack

• Frontend: Next.js 14 (App Router), Tailwind, tRPC.
• Backend: Fastify, tRPC, BullMQ.
• Data: PostgreSQL (pg), Redis, In-Memory Vector Store.
• AI: OpenAI (Embeddings + Chat Completion).

Prerequisites

• Node.js v20+
• pnpm
• Docker (for Postgres/Redis) OR local instances.

Quick Start

1. Install Dependencies

   pnpm install

2. Environment Configuration Copy .env.example to .env and fill in your keys.

   • OPENAI_API_KEY: Required for real AI responses.
   • CLERK_*: Required for auth (or use dev backdoor).

   Minimum .env:

   DATABASE_URL="postgresql://postgres:postgres@localhost:5432/lumina"
   REDIS_URL="redis://localhost:6379"
   API_PORT="3001"
   WEB_URL="http://localhost:3000"

3. Initialize System This creates DB tables and ensures upload directories exist.

   pnpm run db:init

4. Run All Services Starts Web, API, and Worker concurrently.

   pnpm run dev:all

Lite Mode

For resource-constrained environments (like AI Studio Preview), runs in-memory without Postgres/Redis.

pnpm dev:web

Access at http://localhost:3000/chat.