True Cost

True Cost calculated
Input by Manual Entry
Input by Amazon Link

Inspiration

Every time we buy something online, we rarely think about where it came from, how it got to us, or what it cost the planet. We wanted to give full transparency to what it is your buying, from environmental impact to other ethical concerns such as labor laws and public company commitments.

What it does

TrueCost is a "true cost calculator" for consumer products. You paste an Amazon product URL (or describe any product manually), and TrueCost returns:

An environmental grade (A+ to F) with an overall score

CO₂ emissions, water usage, and decomposition time for the product
A labor ethics score based on the brand's supply chain transparency, union recognition, wage disclosure, and modern slavery policies
A climate transparency breakdown (decarbonization, energy, traceability, accountability)
An interactive 3D globe showing the full supply chain route, from raw material origin through manufacturing and distribution to your door, with per-leg CO₂, distance, and transit time

How we built it

Frontend: React + TypeScript + Vite, with Tailwind CSS and react-globe.gl for the 3D supply chain map
Backend: FastAPI (Python), with async parallel data fetching
Rainforest API - scrapes Amazon product data (title, brand, weight, category) from a URL
Climatiq API - calculates CO₂ emissions using matched emission factors by product category and weight
WikiRate + wikirate4py - pulls real corporate sustainability data (Fashion Revolution index, Clean Clothes Campaign scores, gender pay gap, modern slavery statements)
Cohere AI (command-a-03-2025) - generates a structured supply chain route (origin → manufacturing → distribution → destination) with transport modes
Geopy / Nominatim - geocodes supply chain stops and emission factor regions to lat/lng for the globe

Challenges we ran into

Climatiq's emission factor search doesn't always return a match for niche product categories, so we built a category fallback system to map product types to known activity IDs
WikiRate data is sparse for many brands, we had to handle missing metrics gracefully and weight scores only against metrics that actually returned data
Cohere's structured JSON output needed careful schema design to reliably produce valid supply chain paths
Syncing async (FastAPI/httpx) and sync (wikirate4py/geopy) code required running blocking calls in a thread pool executor

Accomplishments that we're proud of

The globe visualization makes the supply chain feel real and tangible in a way a table never could Pulling live, real-world corporate ethics data from WikiRate gives the scores actual credibility rather than just estimates
The pipeline handles both Amazon URLs and fully manual product entry, making it useful beyond just Amazon

What we learned

How to chain multiple third-party APIs with very different data models into a single coherent response The complexity of supply chain emissions accounting, even a simple product has a surprisingly long journey
How sparse and inconsistent public corporate sustainability data still is, even for major brands