The Problem

In an information-rich world, finding specific answers within vast documents (reports, contracts, manuals) is inefficient. Traditional keyword searches lack semantic understanding, leading to missed insights and tedious manual review. Businesses face challenges in securely integrating private data with advanced AI for conversational queries, often resorting to fragmented solutions that lack privacy and multi-user data isolation.

The Solution

MindDoc is a secure, multi-tenant AI Chatbot and AI Agent platform built on Next.js and Supabase. It features a robust ingestion pipeline: documents (PDF, DOCX, TXT) are parsed via officeparser, intelligently chunked, and transformed into vector embeddings using Google Gemini's embedding-001 model. These are stored in Supabase's pg_vector database with associated user IDs. For Q&A, a custom SupabaseUserRetriever ensures RAG LLM queries strictly target the user's specific document, enforced by RLS. LangChain orchestrates conversational memory, feeding context and user questions to Gemini's gemini-2.5-flash-lite-preview-06-17 LLM for accurate, context-aware answers. This creates an AI-powered, conversational interface for private document data.

Tech Stack

Gallery