🚀 FormFlow AI

🚧 Currently under active development — more features coming soon.

Human-in-the-loop OCR-powered form automation system Extract data from images → auto-fill forms → validate in real UI → submit → repeat.

---

✨ Overview

FormFlow AI is designed to automate repetitive data entry workflows while keeping humans in control.

It extracts structured data from images using OCR, processes it into records, and assists users in entering the data into web forms. Users review and correct data directly in the target application before submission.

---

🧠 How It Works

Images → OCR → Parsing → Records Queue → Auto-fill Form → Human Review → Submit → Next Record

---

🔥 Key Features

• 📸 OCR-based data extraction using Tesseract
• 🧩 Structured parsing with normalization and validation
• 🔁 Retry mechanism with logging and error handling
• 🧑 Human-in-the-loop validation inside the actual web application
• 🔗 Config-driven field mapping (no code changes required)
• ⚡ Lightweight and local-first design

---

🧰 Tech Stack

• Python
• Tesseract OCR
• Playwright

---

📁 Project Structure

app/        # OCR, parser, automation
utils/      # logging, file handling
config/     # form and parsing configurations
data/       # input images and runtime data

---

⚙️ Setup & Run

1. Create Virtual Environment

python -m venv venv

Activate:

• Mac/Linux

source venv/bin/activate

• Windows

venv\Scripts\activate

---

2. Install Dependencies

pip install -r requirements.txt
playwright install

---

3. Install Tesseract OCR

• Mac (Homebrew)

brew install tesseract

• Ubuntu / Linux

sudo apt update
sudo apt install tesseract-ocr

• Windows

1. Download installer from: https://github.com/tesseract-ocr/tesseract

2. Install it

3. Add to PATH:

   C:\Program Files\Tesseract-OCR
   

⚠️ Verify installation:

tesseract --version

---

4. Run the Application

python main.py

---

🧪 Example Input

xt1234 john doe california
xt5678 alice smith new york

---

🚀 Use Cases

• Data entry automation
• Back-office workflows
• OCR-based ingestion pipelines
• Form processing systems

---

🛣️ Roadmap

• Streamlit-based review dashboard
• Multi-form support
• AI-assisted parsing fallback
• Batch processing enhancements

---

📜 License

MIT License © 2026 Minhaz Alam

---

👤 Author

Minhaz Alam