🚰 DocuPipe Desktop

A simple, user-friendly desktop app to effortlessly upload and download datasets with DocuPipe.

---

🚀 What is DocuPipe?

DocuPipe is a powerful document processing platform that transforms your documents into searchable, structured data:

• 🔍 Searchable PDFs: Instantly find text using Ctrl+F (Windows/Linux) or Command+F (Mac), thanks to built-in OCR technology.
• 📦 Standardized JSON Data: Extract structured data from your documents based on custom schemas defined at docupipe.ai.

🗂️ Example Output

For instance, from a lease agreement, DocuPipe can extract:

{
  "rentalAmount": 2000,
  "rentalCurrency": "USD",
  "leaseStartDate": "2025-05-01",
  "leaseEndDate": "2026-04-30",
  "petsAllowed": false,
  "securityDeposit": 3000
}

✅ DocuPipe handles checkmarks, handwriting, tables, signatures, and more—if you can see it, DocuPipe can process it.

---

🎯 Features

📤 Easy Upload

• Select "Upload Dataset"
• Choose your folder with documents
• Name your dataset
• (Optional) Choose a schema to standardize the documents (e.g. "rental schema")
• Hit "Confirm Upload"

📥 Simple Download

• Select "Download Dataset Results"
• Pick your dataset from the list
• Choose your save location
• Click "Confirm Download"

Downloading the dataset will give you the following things:

1. Every file is downloaded as a PDF with an OCR layer (that means its searchable and copyable) - even if your original upload was an image ot html. So invoice.png -> invoice.pdf
2. If you chose to also standardize the results, you will also get a .json file with the standardized results. So e.g. rental.pdf -> rental.json

❓ What's a Schema?

• A schema is a set of expectations of what you want to extract from a document. Rental amount in a lease, patient insurance number in a medical record, etc.
• You can create your own schemas at docupipe.ai. Simply pick a few example documents, and explain with words what you want to extract. DocuPipe will then create a schema for you.
• Once a schema is made, you can apply it to documents of varying layouts, formats, and even languages - even if the schema was created using an English document, you can expect it to work out of the box in German, Chinese or any other 60+ supported languages.

---

🛠️ Getting Started

✅ Prerequisites

• Python 3.9+
• DocuPipe account with an API key (Get yours here)

📦 Installation

Option 1: Run from Source

git clone git@github.com:urimerhav/docupipe-desktop.git
cd docupipe-desktop
pip install -e .
flet run src/main.py

Option 2: Build Executable

bash build.sh
# Find your executable in the 'dist' folder

🚦 First Use

• Launch the app
• Enter your API key
• You're set to upload & download datasets!

---

📑 Supported File Types

• 📄 PDF (.pdf)
• 🖼️ Images (.jpg, .jpeg, .png, .tiff, .webp)
• 📝 Text (.txt)
• 📃 HTML and Word (.html, .docx)

---

🐞 Troubleshooting

• ⚠️ Errors: Visible in-app
• 📁 Detailed Logs:
  • macOS: ~/Library/Application Support/DocuPipe/logs/
  • Windows: %LOCALAPPDATA%\DocuPipe\logs\
  • Linux: ~/.docupipe/logs/

---

🙋‍♂️ Need Help?

• Review your logs first
• Reach out to DocuPipe Support with logs for swift assistance

---

📄 License

MIT License