✂️ R.E.D.A.C.T V2

Restricting Exposed Data by Anonymization for Confidential Transmission

🔐 An Electron-based desktop app to redact sensitive PII from text documents and clipboard content using the Hugging Face Inference API, with a preview and restore-friendly workflow.

---

🚀 Features

• 🧠 Redacts PII using Hugging Face's dslim/bert-base-NER model (NER-based)
• ✍️ Inline editing of redacted content
• 📋 Clipboard redaction/restoration with one click
• 📂 Drag-and-drop file support for .txt, .pdf, .doc, .docx
• 🔔 Popup alerts and feedback
• 📄 Generates -REDACTED clones while preserving original format

---

🖼️ Preview

---

🛠️ Installation

1. Clone the Repository

   git clone https://github.com/KarthikBalaji-007/REDACT V2.git
   cd Freshathon

2. Install Dependencies

   npm install

3. Configure Hugging Face API

   • Create a .env file in the root folder
   • Add your API key like this:

     HUGGINGFACE_API_KEY=your_api_key_here

4. Run the App

   npm start

---

🧑‍💻 Usage

📋 Redact Clipboard Content

Click Redact Clipboard → Detected PII is replaced with tags → Displayed in new window

♻️ Restore Clipboard Content

Click Restore Clipboard → Original content restored from mapping → Displayed in new window

📄 Redact a Document

Drag-and-drop .txt, .pdf, .doc, or .docx → Redacted version shown → Saved with -REDACTED suffix

✏️ Edit Redacted Text

Click on redacted tags to manually edit or restore

---

⚠️ Note on Accuracy

This app uses the dslim/bert-base-NER model from Hugging Face’s Inference API for redaction.
It offers good baseline coverage but may occasionally miss certain entities.
Always review output before sharing externally.

---

📁 Examples

Check the examples/ folder for sample input/output:

• example_data.txt – Sample input text
• example_data.txt-REDACTED.txt – Redacted version
• example_data.pdf – Sample PDF
• example_data.pdf-REDACTED.txt – Redacted output in plain text

---

📜 License

MIT License

---

🌟 Feel free to fork, star, and contribute to make R.E.D.A.C.T. even better!