OCR Text Vision Pro

AI-powered OCR application using Llama 3.2 Vision for advanced image understanding and text extraction

✨ Features

📈 General OCR & Content Recognition

• Text Extraction: Extract readable content from any image
• LaTeX Conversion: Convert mathematical equations to LaTeX code with live rendering
• Code Extraction: Extract and format code snippets from screenshots
• Chart Analysis: Describe charts, diagrams, and visual data

📑 Advanced Document Intelligence

• Document VQA: Ask specific questions about document content
• Structured Extraction: Extract invoice numbers, dates, amounts, etc.
• Form Processing: Handle contracts, receipts, and business documents

❓ Intelligent Visual Question Answering

• Scene Understanding: Analyze and describe image content
• Object Recognition: Identify and reason about objects in images
• Visual Reasoning: Answer complex questions about visual content

🗣️ Multi-modal Chat Assistant

• Interactive Conversations: Chat with AI about uploaded images
• Context Awareness: Maintains conversation history
• Real-time Responses: Instant AI-powered image analysis

🔁 Workflow

flowchart TB
 subgraph subGraph0["Streamlit App"]
        UI["Presentation Layer (Streamlit UI)"]
        PIL["Image Preprocessor (PIL)"]
        APIClient["API Client (requests)"]
        Session["Session Manager<br>(in-memory API key &amp; chat history)"]
  end
    Browser["User’s Web Browser"] -- UI event / upload image --> UI
    UI -- image --> PIL
    PIL -- processed image --> APIClient
    UI -- store API key & history --> Session
    Session -- provide API key --> APIClient
    APIClient -- HTTP POST --> External["OpenRouter API<br>(Llama 3.2 Vision)"]
    External -- JSON response --> APIClient
    APIClient -- parsed results --> UI
    Deployment["Deployment Environment<br>(Streamlit Cloud or Docker)"] -- hosts --> UI

     UI:::frontend
     PIL:::app
     APIClient:::app
     Session:::app
     Browser:::frontend
     External:::external
     Deployment:::deployment
    classDef frontend fill:#D6EAF8,stroke:#1B4F72
    classDef app fill:#D5F5E3,stroke:#145A32
    classDef external fill:#FAD7A0,stroke:#B9770E
    classDef deployment fill:#E5E7E9,stroke:#566573
    style Browser color:#000000
    style External color:#000000
    style Deployment color:#000000
    click UI "https://github.com/bcastelino/ocr-text-vision-pro/blob/main/ocr_app.py"
    click PIL "https://github.com/bcastelino/ocr-text-vision-pro/blob/main/ocr_app.py"
    click APIClient "https://github.com/bcastelino/ocr-text-vision-pro/blob/main/ocr_app.py"
    click Session "https://github.com/bcastelino/ocr-text-vision-pro/blob/main/ocr_app.py"

Loading

🚀 Quick Start

Option 1: Streamlit Community Cloud (Recommended)

1. Deploy directly - No setup required!
2. Enter your OpenRouter API key (free)
3. Start uploading images and extracting text!

Option 2: Local Development

# Clone the repository
git clone https://github.com/bcastelino/ocr-text-vision-pro.git
cd ocr-text-vision-pro

# Install dependencies
pip install -r requirements.txt

# Run the application
streamlit run ocr_app.py

Option 3: Docker

# Build and run with Docker
docker build -t ocr-text-vision-pro .
docker run -p 8501:8501 ocr-text-vision-pro

🔧 Tech Stack

• Frontend: Streamlit (Python web framework)
• AI Model: Llama 3.2-11B Vision (via OpenRouter API)
• Image Processing: PIL/Pillow
• HTTP Client: Requests
• Deployment: Streamlit Community Cloud

📋 Requirements

• Python 3.11+
• OpenRouter API key (free tier available)
• Modern web browser

🎯 Use Cases

• Students: Extract text from lecture slides and handwritten notes
• Researchers: Convert mathematical equations to LaTeX
• Developers: Extract code from screenshots and documentation
• Business: Process invoices, receipts, and contracts
• Content Creators: Analyze charts and extract data for reports

🔐 Privacy & Security

• API keys are stored only in session (not persisted)
• No image data is stored on my server
• All processing happens through the secure OpenRouter API
• Runs entirely in your browser session

🤝 Contributing

Welcome contributors! Please feel free to submit issues and enhancement requests.

📄 License

This project is open source and available under the MIT License.

🐱‍👤 Author

Brian Denis Castelino Data Analytics Engineer | AI Enthusiast I turn vague ideas into clean, working systems, because someone’s got to 🤖

---

Made with ❤️ using Streamlit and Llama 3.2 Vision