debrisai

Debris AI

Debris AI is a hackathon project that analyzes images of construction debris to identify materials, estimate reuse feasibility, and provide safe, non-structural reuse suggestions.

Project Structure

The project is organized into the following directories:

Backend

The backend is a Python Flask application designed to be deployed on Cloud Run. It exposes two endpoints:

The backend also integrates with Confluent Cloud to stream events for every request.

Safety

Debris AI prioritizes safety by providing a disclaimer with every analysis and offering conservative, non-structural reuse suggestions. A verification layer is also in place to re-prompt the model with stricter instructions if the initial confidence score is low.

Getting Started

To get started with the project, please refer to the documentation in the deployment/ directory.