## Inspiration:- While viewing a documentary film on the Environmental Impact, a haunting thought crossed my mind: do we really need everything we buy, or is everything just for the sake of fashion and showoff? I realized that when we buy something, like a pair of leather shoes, we are not just paying the price tag, but unknowingly contributing to the destruction of our environment. I realized that innocent animals are being slaughtered in slaughterhouses, and carcinogens are being used in leather factories, destroying our environment. The worst part is that slum dwellers are forced to work with these chemicals, causing irreparable damage to their health.

I realized that everything we buy causes a ripple effect of pain in the world, which has severe short- and long-term effects, but we are not even aware of them. This gave me an idea: what if we, as consumers, knew the real cost of what we buy? What if there was a way of guiding people towards buying products like Bio-enzymes, which are not harmful, rather than carcinogens?That is why I created GhostCost-AI, an intelligent agent that audits the manufacturing, environmental, and human impact of products to show consumers the true cost and debt to society.

What it does

GhostCost-AI is a transparent mirror to consumers. Users can scan any product or search to get a complete audit on four essential pillars: Manufacturing, Human Impact, Animal Cost, and Environmental Debt. The AI agent calculates a percentage-based Ghost Score to reflect how toxic a product is to the environment. GhostCost-AI also offers consumers green alternatives and solutions to encourage consumers to practice conscious shopping.

How I built it

As a first-year computer science student, I created GhostCost-AI by using Python and Streamlit to design the front end. I also integrated Google’s Gemini 1.5/2.5 Flash model using the Stable V1 SDK to ensure reliable and agentic reasoning. Additionally, I used Session State to enable users to have a chatbot session to discuss conscious shopping in real time.

Challenges I ran into

The journey was not an easy one. Technical challenges abounded. Technical challenges included 404 NOT_FOUND versioning and 429 RESOURCE_EXHAUSTED rate limits in the last sprint. I had to learn to manage API versioning on the fly. Eventually, I had to re-provision the infrastructure on Google Cloud and migrate from legacy beta endpoints to stable production SDK to achieve 100% uptime.

Accomplishments that I'm proud of

I am proud to say that I was able to transform 96% error rates into a fully functional AI Auditor in just 48 hours. Being able to master the transition to Stable V1 SDK and create a product that aligns with my objective of developing technology that is not only ethical but also impactful is a huge milestone in my journey to becoming an aspiring businesswoman.

What I learned

I learned that being a coder is not only about coding but also requires technical resilience. I learned so much more about Multimodal AI, Cloud API orchestration, and the critical need for transparency in global supply chains.

What's next for GhostCost AI : Global Ethical Auditor

Built With

  • genai
  • google-gemini-1.5/2.5-flash-api
  • pillow
  • python
  • stable
  • streamlit
  • v1
Share this project:

Updates