NassaQ - Document Digitization Platform¶

What is NassaQ?¶

NassaQ (Arabic: نسّـق, meaning "to organize" or "to format") is an AI-powered document management and digitization platform. It enables organizations to upload physical or digital documents and have them automatically processed through intelligent OCR pipelines that support both Arabic and English text extraction.

The platform transforms unstructured documents into searchable, categorized, and retrievable digital assets -- making it possible to go from a scanned page to a fully indexed, queryable piece of content in seconds.

Core Capabilities¶

• Dual-Engine OCR -- A smart routing system that selects between PaddleOCR (optimized for speed and Latin scripts) and EasyOCR (optimized for Arabic cursive text) based on detected language
• Multi-Format Processing -- Handles PDFs (both digital and scanned), images (JPEG, PNG), and plain text files
• Asynchronous Processing -- Documents are queued via RabbitMQ and processed in the background, so users never wait
• Cloud-Native Storage -- All files are stored in Azure Blob Storage with metadata tracked in Azure SQL Server
• Bilingual Interface -- A fully localized web application supporting English and Arabic (RTL) out of the box
• Role-Based Access Control -- Users, roles, and permissions are managed through a structured authorization system

Platform Architecture at a Glance¶

NassaQ follows a microservices architecture with three independently deployable components:

graph LR
    A["<b>User Interface</b><br/>React + TypeScript<br/>Port 8080"] -->|REST API| B["<b>Backend Server</b><br/>FastAPI + SQLAlchemy<br/>Port 8000"]
    B -->|AMQP| C[("<b>RabbitMQ</b><br/>Message Broker<br/>Port 5672")]
    C -->|Consume| D["<b>OCR Worker</b><br/>FastAPI + PaddleOCR<br/>Port 8001"]
    B -->|Read/Write| E[("Azure SQL<br/>Server")]
    B -->|Upload/Download| F[("Azure Blob<br/>Storage")]
    D -->|Read/Write| E
    D -->|Download| F
    B -.->|Planned| G[("Azure Cosmos DB<br/>MongoDB")]

Quick Start¶

Get the entire platform running locally with Docker Compose.

Prerequisites¶

• Docker and Docker Compose installed
• Azure credentials for SQL Server, Blob Storage, and (optionally) Cosmos DB
• A copy of each repository cloned into a shared parent directory

1. Clone the Repositories¶

mkdir nassaq && cd nassaq

git clone git@github.com:NassaQ/server.git
git clone git@github.com:NassaQ/ocr-api.git ocr
git clone git@github.com:NassaQ/User_Interface.git frontend

2. Configure Environment Variables¶

Each backend service requires its own .env file. Copy the examples and fill in your Azure credentials:

cp server/.env.example server/.env
cp ocr/.env.example ocr/.env

Each .env.example file documents every required and optional variable. See Backend Server Setup and OCR API Setup for descriptions of each variable.

3. Launch with Docker Compose¶

docker compose up --build

This starts three containers:

4. Start the Frontend¶

The frontend runs separately (not yet containerized):

cd frontend
npm install    # or: bun install
npm run dev    # or: bun dev

The UI will be available at http://localhost:8080.

How It Works¶

The document processing flow follows these steps:

sequenceDiagram
    actor User
    participant UI as User Interface
    participant API as Backend Server
    participant Blob as Azure Blob Storage
    participant DB as Azure SQL Server
    participant MQ as RabbitMQ
    participant OCR as OCR Worker

    User->>UI: Upload document
    UI->>API: POST /api/v1/docs/upload
    API->>Blob: Store original file
    API->>DB: Create Document record<br/>(status: Queued)
    API->>MQ: Publish to ocr_queue
    API-->>UI: 200 OK (doc_id)

    MQ->>OCR: Deliver message
    OCR->>DB: Update status: Processing
    OCR->>Blob: Download file
    OCR->>OCR: Run smart OCR pipeline
    OCR->>DB: Update status: Finished
    OCR-->>MQ: Acknowledge message

    User->>UI: Check status
    UI->>API: GET /api/v1/docs/{id}/status
    API->>DB: Query ProcessingStatus
    API-->>UI: { status: "Finished" }

Technology Stack¶

Backend Services (Python)¶

Frontend (TypeScript)¶

Infrastructure¶