MIT-AI-Hackathon-2025

AI-Powered Personal CRM Assistant

An AI-powered assistant that extracts contact details from business cards, images, and PDFs, organizes them into a searchable database, and sends personalized LinkedIn connection requests.

🚀 Features

• 📤 Upload business cards, PDFs, or photos to extract contact info
• 🧠 NLP and OCR pipeline using Tesseract and spaCy
• 🗃️ Store and search contacts via SQLite
• 🔗 Send LinkedIn connection requests with custom messages (via Selenium)
• 🖥️ Lightweight frontend with drag-and-drop UI

🧰 Tech Stack

• Python, Flask – API and server
• Tesseract OCR – Extract text from images
• spaCy – Named Entity Recognition for contacts
• SQLite – Contact database
• Selenium + ChromeDriver – LinkedIn automation
• HTML/CSS/JS – Frontend
• Docker – Containerization

📦 Installation

Option 1: Run with Docker

git clone https://github.com/your-username/personal-crm-assistant.git
cd personal-crm-assistant
docker build -t personal-crm .
docker run -p 5000:5000 personal-crm

Option 2: Run locally

pip install -r requirements.txt
python -m spacy download en_core_web_sm
python personal_crm_assistant.py

🔍 API Endpoints

• POST /upload – Upload a file for contact extraction
• GET /search?q=John – Search contacts
• POST /linkedin/connect – Send LinkedIn request (requires logged-in Chrome session)

🖼️ Frontend Usage

Open index.html in your browser. Make sure the backend is running on localhost:5000.

⚠️ Limitations

• LinkedIn automation requires manual login into the browser.
• Running Selenium inside Docker may require GUI or headless setup.
• Not suitable for production without HTTPS and authentication.

📌 To-Do

• ✅ Improve OCR accuracy with layout parsing
• 🛠️ Batch processing for Excel/CSV
• 🌐 Chrome Extension integration

🧠 Credits

Built as a 12-hour project using open-source tools. Contributions welcome!