IDP - Intelligent Document Processing

A full-stack platform for intelligent document processing with OCR and AI.

Features

• Template Management: Create extraction templates with AI-powered field suggestions
• Document Upload: Upload PDFs and images with automatic data extraction
• Auto-Classification: Automatic document classification into existing templates
• Multi-Provider AI: Support for OpenAI, Claude (Anthropic), and Gemini (Google)
• Extraction Review: Review interface with inline editing and confidence badges
• Data Tables: Tabular view of extracted data by template
• CSV Export: Export extracted data to CSV
• Activity Log: Full history of all platform operations
• Batch Upload: Upload multiple documents at once
• Connection Testing: Test AI and OCR connections directly from the UI
• Dashboard: Processing statistics and overview
• Template Assignment: Manual template assignment for unclassified documents

Tech Stack

Layer	Technology
Backend	FastAPI + SQLAlchemy + SQLite
OCR	pdfplumber (primary) + Tesseract + pdf2image (fallback)
AI	OpenAI / Claude / Gemini (configurable)
Frontend	Next.js + TypeScript + Tailwind CSS + shadcn/ui
State	TanStack Query (React Query)

Prerequisites

• Python 3.10+
• Node.js 18+
• Tesseract OCR (optional, only needed for scanned PDFs)
• libmagic (file type detection)

Installing system dependencies

macOS

brew install python3 node tesseract libmagic poppler

Ubuntu / Debian

sudo apt install python3 python3-pip python3-venv nodejs npm tesseract-ocr libmagic1 poppler-utils

Fedora

sudo dnf install python3 python3-pip nodejs npm tesseract libmagic poppler-utils

Windows

1. Python: https://www.python.org/downloads/
2. Node.js: https://nodejs.org/
3. Tesseract: https://github.com/UB-Mannheim/tesseract/wiki
4. Poppler: https://github.com/oschwartz10612/poppler-windows/releases

Quick Start

The project includes a Makefile that auto-detects your toolchain (Python, Node, OS) and handles dependency installation automatically.

# Full setup (check dependencies, create .env, install packages)
make setup

# Start both backend and frontend in dev mode (with hot-reload)
make dev

That's it. The make dev command will automatically install missing dependencies before starting.

Available Make Commands

Run make help to see all commands with your detected toolchain:

Command	Description
make setup	Full project setup (check deps, create .env, install)
make install	Install all dependencies
make dev	Start backend + frontend in dev mode (parallel, hot-reload)
make dev-backend	Start only the backend in dev mode
make dev-frontend	Start only the frontend in dev mode
make start	Start both in production mode
make build	Build frontend for production
make test	Run backend tests
make lint	Run frontend linter
make migrate	Run database migrations
make migrate-new MSG="desc"	Create a new migration
make venv	Create a Python virtual environment (.venv)
make env	Create .env from .env.example
make clean	Remove build artifacts and caches
make check-deps	Verify all required tools are installed

Using a virtual environment (recommended)

make venv
source .venv/bin/activate   # Linux/macOS
# .venv\Scripts\activate    # Windows
make setup
make dev

Manual Setup

If you prefer not to use Make:

Backend

cd backend
pip install -r requirements.txt
cp .env.example .env
# Edit .env with your API key (or configure via the UI)
python run.py

Frontend

cd frontend
npm install
npm run dev

Usage

1. Initial Configuration

• Open http://localhost:3000
• Go to Settings and configure the AI provider + API key
• Click "Test AI Connection" to verify

2. Create a Template

• Go to Templates > New Template
• Enter a name (e.g., "Invoice", "Contract", "Receipt")
• Upload a sample document
• AI will automatically suggest extractable fields
• Review, add, or remove fields as needed

3. Process Documents

• Go to Upload and select one or more documents
• Choose a template or let the AI classify automatically
• The system runs: OCR > Classification > Extraction > Storage

4. Review Data

• In Documents, click on a processed document
• View the document side by side with extracted data
• Edit incorrect values and approve the extraction

5. View & Export

• In Extracted Data, see all data in tabular format
• Filter by template, search for specific values
• Export to CSV for use in spreadsheets

API Documentation

Swagger UI: http://localhost:8000/docs

Main Endpoints

Method	Endpoint	Description
GET	/api/dashboard/stats	General statistics
POST	/api/templates	Create template
POST	/api/templates/{id}/suggest-fields	AI field suggestions
POST	/api/documents/upload	Upload document
POST	/api/documents/upload-batch	Batch upload
GET	/api/documents/{id}	Document details
PUT	/api/documents/{id}/assign-template	Assign template
GET	/api/data/templates/{id}/table	Tabular data
GET	/api/data/templates/{id}/export	Export CSV
GET	/api/activity	Activity log
POST	/api/settings/test-ai	Test AI connection
POST	/api/settings/test-ocr	Test OCR