Blueprint Intelligence Engine

AI-powered blueprint parsing engine that extracts room layouts and dimensions from residential floor plans, then calculates material quantities and cost estimates.

Overview

The Blueprint Intelligence Engine (BIE) is a tool designed to:

1. Accept residential floor plan uploads (PDF/image formats)
2. Parse blueprints using GPT-4 Vision to identify rooms and dimensions
3. Calculate material estimates for flooring, drywall, and paint
4. Display results with cost ranges (low/mid/high)

This project has two potential products:

• Consumer App: A web application for homeowners and contractors
• White-Label API: A B2B API product for integration into other platforms

Project Status

Current Phase: Phase 0 - Setup & Foundation

Phase	Status	Description
Phase 0	🟡 In Progress	Setup & Foundation
Phase 1	⚪ Not Started	Proof of Concept
Phase 2	⚪ Not Started	MVP Development
Phase 3	⚪ Not Started	Launch & Iterate
Phase 4	⚪ Not Started	White-Label API

Project Structure

blueprint-intelligence-engine/
├── data/
│   ├── blueprints/          # Raw blueprint files (PDF, PNG, JPG)
│   ├── processed/           # Pre-processed images
│   └── ground_truth/        # Manual annotations for testing
├── src/
│   ├── preprocessing/       # Image preprocessing modules
│   ├── parsing/             # AI parsing logic
│   ├── calculation/         # Material calculation engine
│   └── api/                 # API wrapper (Phase 4)
├── tests/                   # Test scripts
├── notebooks/               # Jupyter notebooks for experimentation
├── docs/                    # Documentation
├── .env.example             # Example environment variables
├── .gitignore
├── requirements.txt
└── README.md

Getting Started

Prerequisites

• Python 3.11+
• Tesseract OCR
• Poppler (for PDF processing)
• OpenAI API key

Installation

1. Clone the repository:

   git clone https://github.com/scruffyjerk/blueprint-intelligence-engine.git
   cd blueprint-intelligence-engine

2. Create and activate a virtual environment:

   python3 -m venv venv
   source venv/bin/activate  # On Windows: venv\Scripts\activate

3. Install dependencies:

   pip install -r requirements.txt

4. Install system dependencies:

   # Ubuntu/Debian
   sudo apt-get install tesseract-ocr poppler-utils
   
   # macOS
   brew install tesseract poppler

5. Configure environment variables:

   cp .env.example .env
   # Edit .env and add your OpenAI API key

Running the Pipeline

python src/poc_pipeline.py --input data/blueprints/sample.pdf

Technical Stack

Component	Technology
AI Model	GPT-4 Vision (gpt-4o / gpt-4.1-mini)
OCR	Tesseract
Image Processing	OpenCV, Pillow
PDF Processing	pdf2image, Poppler
Backend (MVP)	FastAPI
Frontend (MVP)	React

Success Metrics

Metric	PoC Target	MVP Target
Room Detection Accuracy	70%	85%
Dimension Extraction Accuracy	50%	70%
Processing Time	< 60s	< 30s
Cost Per Blueprint	< $0.15	< $0.10

Contributing

This is currently a private project in development. Contribution guidelines will be added when the project opens for collaboration.

License

TBD

Acknowledgments

Built with assistance from AI tools including Manus, Claude, and GitHub Copilot.

Trigger redeploy Tue Jan 20 18:26:28 EST 2026