README.md

---
license: mit
title: Future Earth - Agentic Eco Advisor
sdk: gradio
emoji: 🚀
colorFrom: green
colorTo: green
pinned: false
short_description: Instant environmental assessment before purchasing products
tags:
- mcp-in-action-track-customer
- mcp-in-action-track-creative
sdk_version: 6.0.1
Team_member: Kathir1311
---
# 🌍 Future Earth - Agentic Eco Advisor
 
> **AI-Powered Environmental Impact Assessment Agent** | Built for Future Earth

## 🎥 Demo Video
[![Watch the Demo](https://img.youtube.com/vi/EQ6RWTNWkq0/hqdefault.jpg)](https://www.youtube.com/watch?v=EQ6RWTNWkq0)

## 🔗 Social Media Post Links
- LinkedIn Launch Post: https://www.linkedin.com/posts/shyamchandar-d-30447385_futureearth-huggingface-mcp-activity-7400088484062875648-Zrpt?utm_source=share&utm_medium=member_desktop&rcm=ACoAABILTPQBvR7C68QDSz999ddmxHHC99I0vwI

- X Post - https://x.com/Dshyam03/status/1994326237649453219?s=20

 
An autonomous AI agent that analyzes products for environmental impact using **computer vision**, **lifecycle assessment**, and **RAG** (Retrieval Augmented Generation) LlamaIndex to provide instant sustainability recommendations.
 
---
## <u>Problem Statement</u>
 
### **We're Destroying Our Planet, One Purchase at a Time**

 **The Hard Truth:**

- **Global carbon emissions increased by 60%** in the last 30 years
- **E-waste is growing 5x faster** than it's being recycled
- **Only 9% of plastic** ever produced has been recycled
- **Fashion industry alone** accounts for 10% of global emissions
- **Each year, our consumer choices damage Earth by ~2.5%** of total environmental capacity
 
### **If you care about this Earth and want to save it for future generations, but don't know how...**
 
**The Problem is:**

Most people **want** to make eco-friendly choices, but:

- They don't have **instant access** to environmental impact data while shopping
- Marketing claims are filled with **greenwashing** and false promises
- Carbon footprint calculations are **too complex** for average consumers
- There's **no easy way** to identify sustainable alternatives on-the-go
 
### **Imagine this scenario:**

You're at a store, holding a product. You want to know:

- "Is this product destroying the planet?"
- "How much carbon footprint does it have?"
- "Is there a better alternative?"
 
**But you can't answer these questions in real-time.** So you buy it anyway.
 
---
 
## <u>The Solution: Future Earth</u>
 
**What if you could point your phone camera at ANY product and instantly know its environmental impact?**
 
Future Earth is an **AI-powered agent** that:
- **Sees** your product through computer vision
- **Thinks** through autonomous reasoning and lifecycle analysis
- **Recommends** better alternatives with quantified impact scores

## <u>How to use Future Earth</u>
 - **Step 1** Launch our Huggingface webapp URL
 - **Step 2** Scan or Upload any product image (examle: Leather Shoes, Plastic watter bottle, bags, Playstation,....etc)
 - **Step 3** Click Analyze Environmental Impact button
 - **You will be getting**
    - Environmental impact score
    - Carbon foot print
    - Environmental Issues
    - Material Analysis
    - Lifecycle Insights
    - **Better Alternatives**

### **If we all start making informed choices today, we can:**
- **Reduce global carbon footprint by 20-30%** through conscious consumption
- **Cut plastic waste in half** by choosing reusable alternatives
- **Extend product lifecycles** by 3-5x through better purchasing decisions
- **Save our Earth for future generations** one product at a time
 
### **The Power of Collective Action**

## 🛠️ Technology Stack

### Core AI & Language Models
- **Claude Sonnet 4 (2025-05-14)** - Anthropic's latest multimodal AI model
  - Vision analysis for product identification
  - Environmental reasoning and assessment generation
  - Multi-step autonomous agent planning

### Framework & Infrastructure
- **Gradio 6** - Modern web UI framework with real-time streaming
  - Mobile-responsive design
  - Progress tracking and status updates
  - Camera integration for product scanning

### Agent Architecture
- **Model Context Protocol (MCP)** - Distributed AI computation
  - **MCPVisionServer** - Image analysis and greenwashing detection
  - **MCPReasoningServer** - Lifecycle impact calculations
  - **MCPParsingServer** - Environmental report parsing

### RAG (Retrieval Augmented Generation)
- **LlamaIndex** - Advanced semantic search framework
  - Vector-based knowledge retrieval
  - HuggingFace embeddings (BAAI/bge-small-en-v1.5)
  - 15-category comprehensive environmental knowledge base