Greenwashers!

• 

  Front end UI

• 

  Front end UI

• 

  Company analysis

• 

  Analysis Complete UI

• 

  Peer Assessment Score, Confidence rating, analysis

Inspiration - The idea came from a simple observation. One of us came across a major brand advertising itself as "100% sustainable" and we started asking, what does that actually mean, and is anyone verifying it? That question became the foundation of our project.

What it does - GreenWashers cross references what companies publicly claim against real emissions data, regulatory filings, and third party disclosures, and surfaces where the inconsistencies are. It generates a greenwashing risk score with a confidence rating, a peer assessment score comparing the company against others in the same sector, and cited sources backing every finding.

How we built it - We built a FastAPI backend to handle our core logic, with Actian VectorAI DB over PostgreSQL managing vector storage and similarity search across ESG documents and known greenwashing patterns. We used OpenAI text-embedding-3-large to generate embeddings and power our RAG based report generation, and an OpenAI live search workflow to actively scrape and ingest fresh sources at query time. Everything is served through a single page frontend that handles both the analyze and file upload flows.

Challenges we ran into - The biggest challenge was the inconsistency of environmental data. Sustainability reporting varies significantly across industries, which made building a reliable comparison framework more difficult than we anticipated. Getting our vector search and RAG pipeline to produce consistent, accurate results across that messy data was something we spent a lot of time on.

Accomplishments that we're proud of - We are proud that we got a fully functional end to end pipeline working over a single weekend. From live source scraping to vector search to generated risk reports with citations, it actually works. We are also proud of the peer assessment feature, which adds a layer of context that we think makes the results genuinely useful.

What we learned - We learned a lot about ESG frameworks, vector search, and retrieval augmented generation. We also learned just how unregulated corporate sustainability reporting actually is, which honestly made us feel like this problem is more important than we initially thought.

What's next for Greenwashers! - We want to expand our hardcoded dataset to cover more major companies, improve the accuracy of our risk scoring model, and build out a public facing dashboard where users can browse and compare companies over time. Long term, we think there is a real opportunity to make this a trusted resource for consumers, journalists, and investors alike.

Built With

• actiandb
• data
• fastapi
• html
• javascript
• llm
• openai
• perplexity
• postgressql
• python
• rag
• railway
• three.js
• vector