BarchScan

BarchScan is a modern SaaS platform designed to digitize physical structured documents (logbooks, sheets, registers) into secure, encrypted, and analyzable digital formats.

1. Core Platform Overview

BarchScan combines mobile-optimized capture interfaces with Google Gemini 2.5 Flash vision AI to extract structured tables from page photographs. The system follows a human-in-the-loop workflow:

1. Scan: Users upload or take a photograph of any physical record page.
2. Suggest: The AI model analyzes the image and suggests a structured table.
3. Edit: Users review and refine the suggested table (insert, rename, delete rows/columns; edit cell values).
4. Create: Users save the verified, encrypted data to the cloud.

---

2. Product Tiers & SaaS Model

BarchScan is structured as a two-tier subscription service:

Tier 1: Records Vault ($8/month)

• Goal: Seamless retrieval, digitizing, and long-term secure archiving.
• Core features:
  • Ingest and digitize physical records via mobile-optimized camera capture or file upload.
  • Schema context tracking (learns the structure from previous records to ensure consistency).
  • Guided human-in-the-loop table editing.
  • AES-256 end-to-end data encryption.
  • Session isolation and security.
  • Export options to CSV and JSON formats.

Tier 2: Intelligence Engine ($23/month)

• Goal: Advanced analysis, statistical evaluation, and data intelligence.
• Core features:
  • All Tier 1 Records Vault functionality.
  • Manual and AI-assisted analytics tools.
  • External dataset upload (CSV, Excel, JSON) for combined analysis.
  • Tag and reference (@mention) system to isolate specific data inside chat.
  • Natural language AI Chat Assistant for queries, summaries, and calculations.
  • Automated chart and graph generation.
  • Real-time trend and anomaly detection.

---

3. Project Architecture

BarchScan is built with a decentralized, component-based frontend and serverless API handlers on Vercel:

                  ┌──────────────────────────────────────────────┐
                  │                 USER BROWSER                 │
                  └──────┬───────────────┬────────────────┬──────┘
                         │               │                │
                         ▼               ▼                ▼
                     ┌───────┐      ┌─────────┐      ┌─────────┐
                     │   /   │      │ /upload │      │  /data  │
                     │ Landing  │      │ Capture │      │ Vault & │
                     │ Page  │      │ Ingest  │      │ Engine  │
                     └───────┘      └─────────┘      └─────────┘
                         │               │                │
                         └───────────────┼────────────────┘
                                         ▼ (HTTPS API Calls)
                  ┌──────────────────────────────────────────────┐
                  │                  API LAYER                   │
                  │                 (/api/*)                     │
                  └──────────────────────┬───────────────────────┘
                                         ▼
                 ┌──────────────────────────────────────────────┐
                 │              BACKEND INTEGRATIONS            │
                 │                                              │
                 │ • Google Gemini (Vision & Analytics AI)      │
                 │ • Supabase (Database, Auth, AES Encryption) │
                 │ • Cloudinary (Encrypted Image Store)         │
                 └──────────────────────────────────────────────┘

Directories & Structure

• product-page/: Vite + React project for the marketing landing page (located at / in production).
• public/: Root static output directory.
  • /upload: Legacy ingestion app bundle.
  • /data: Digital vault and analytics engine app bundle.
• api/: Serverless functions (Node.js) handling database operations, AI requests, and uploads.
• docs/: Product planning, business plans, system design documents, and PDF generation scripts.

---

4. Local Development

Prerequisites

• Node.js (v18 or higher)
• Vercel CLI (optional, for api simulation)

Installation

Clone the repository and install dependencies in the root:

npm install

Install dependencies for the product page:

cd product-page
npm install
cd ..

Running Locally

To run the product landing page:

npm run dev

To build and compile all packages to production assets (copies build artifacts to public/):

node deploy_product.js

---

5. Deployment

The project is configured for deployment on the Vercel serverless platform.

Deploying production build:

node deploy_product.js
npx vercel --prod