InsightDocs

InsightDocs is an AI-powered document intelligence platform that lets users upload PDFs, process their contents, search through them semantically, and ask questions in natural language.

The project is built as a full-stack SaaS-style application with a React frontend, an Express backend, MongoDB persistence, and Gemini-powered AI features.

Features

• Secure authentication with JWT
• Protected dashboard experience
• PDF upload and storage
• PDF text extraction and processing
• Text chunking for downstream retrieval
• Embedding generation for semantic search
• Semantic search over document chunks
• AI chat for document Q&A
• Persistent chat history per document
• User profile management
• Production-ready deployment configuration for Render and Vercel

Tech Stack

Frontend

• React
• Vite
• Tailwind CSS
• React Router
• Axios

Backend

• Node.js
• Express.js
• MongoDB + Mongoose
• JWT + bcrypt
• Multer
• Google Gemini API

Deployment

• Render for the backend
• Vercel for the frontend
• MongoDB Atlas for data persistence

Project Architecture

flowchart TB
    Client[React client\nAuthentication · Dashboard · Upload · Chat · History · Profile]
    Client -->|HTTPS / REST API| API[Express API]
    subgraph Services[Backend services]
        Auth[Auth service\nRegister · Login · JWT]
        User[User service\nProfile · Settings]
        Documents[Document service\nUpload · Extract text · Metadata]
        RAG[RAG service\nChunk · Embed · Retrieve]
        Chat[Chat service\nConversation · Answer · History]
    end
    API --> Auth
    API --> User
    API --> Documents
    API --> RAG
    API --> Chat
    Documents --> RAG --> Chat
    Chat --> Gemini[Google Gemini API]
    RAG --> Embeddings[Gemini embedding model]
    Auth --> Mongo[(MongoDB)]
    User --> Mongo
    Documents --> Mongo
    RAG --> Mongo
    Chat --> Mongo

Loading

The dashboard runs the document workflow after upload: upload → text extraction → chunking → embeddings. This keeps each document ready for semantic search and chat as soon as it appears in the document list.

Folder Structure

InsightDocs/
├── backend/
│   ├── src/
│   │   ├── app.js
│   │   ├── server.js
│   │   ├── config/
│   │   ├── controllers/
│   │   ├── middleware/
│   │   ├── models/
│   │   └── routes/
│   ├── test/
│   ├── package.json
│   └── .env.example
├── frontend/
│   ├── src/
│   │   ├── api/
│   │   ├── components/
│   │   ├── pages/
│   │   └── App.jsx
│   ├── package.json
│   └── vite.config.js
├── DEPLOYMENT.md
├── project_plan.md
└── changelog.md

Installation Guide

Prerequisites

• Node.js 18+
• npm 9+
• MongoDB Atlas account
• Google Gemini API key

Backend

cd backend
npm install
cp .env.example .env
npm run dev

Frontend

cd frontend
npm install
npm run dev

Environment Variables

Create a .env file in the backend using the values from .env.example.

Required variables:

• PORT
• NODE_ENV
• CLIENT_URL
• MONGODB_URI
• JWT_SECRET
• JWT_EXPIRES_IN
• GEMINI_API_KEY

API Documentation

Authentication

• POST /api/auth/register
• POST /api/auth/login
• GET /api/auth/me
• POST /api/auth/logout

Documents

• POST /api/documents
• GET /api/documents
• POST /api/documents/:documentId/process
• POST /api/documents/:documentId/chunk
• POST /api/documents/:documentId/embed
• POST /api/documents/:documentId/search
• POST /api/documents/:documentId/chat
• GET /api/documents/:documentId/chat-history

Profile

• GET /api/profile
• PUT /api/profile

Health

• GET /api/health

Screenshots

Placeholder screenshots will be added here after deployment:

• Dashboard overview
• PDF upload experience
• AI chat experience
• Profile settings

Future Improvements

• Add file deletion and document management workflows
• Add richer chat history controls such as delete and rename
• Improve chunking and retrieval quality
• Add support for multi-document conversation context
• Introduce streaming AI responses
• Improve observability and analytics

Contributing

Contributions are welcome.

1. Fork the repository
2. Create a feature branch
3. Commit your changes
4. Open a pull request

License

This project is licensed under the MIT License.