Developed an open-source, production-ready Document Intelligence platform that enables users to have natural language conversations with their documents (PDFs, DOCX, TXT, and images). The system leverages Retrieval-Augmented Generation (RAG) and Semantic Search to provide context-aware responses with source citations. Key architectural features include: ⭐️ Multi-Provider AI Integration: Seamlessly supports OpenRouter (100+ models), OpenAI, and Anthropic. ⭐️ Advanced Document Processing: Integrated IBM’s Docling for high-fidelity table extraction and layout preservation. ⭐️ Vector Search & Storage: Built using Pinecone and pgvector for efficient semantic similarity matching. ⭐️ Scalable Architecture: Implemented background job processing with Inngest for asynchronous document vectorization and OCR. ⭐️ Enterprise-Ready: Features Clerk authentication, multi-tenant organization isolation, and Stripe-integrated billing for SaaS scalability. Modern Tech Stack: Built with Next.js 15, React 19, TypeScript, Prisma ORM, and Tailwind CSS.