ReceiptReveal

Inspiration

Climate concerns are rising as over 70% of people want to reduce their carbon footprint but don’t know where to begin. We wanted to turn everyday behaviour into meaningful environmental impact. Our goal is to make carbon tracking as simple as snapping a photo and as engaging as a game. ReceiptReveal empowers users to understand their carbon footprint through AI and inspires real change through community, rewards, and habit-building.

What it does

ReceiptReveal is an AI-powered sustainability app that lets users upload receipts or log daily activities to instantly estimate their carbon emissions. The app uses Google’s Gemini to analyze text and images, calculates footprint data using LangChain along with a carbon dataset, and presents the result in a gamified dashboard. Users earn Coins for reducing emissions and could redeem points at eco-conscious partner brands.

How we built it

We built ReceiptReveal using a blend of modern web technologies and advanced AI frameworks to create a seamless, intelligent, and interactive sustainability platform. The frontend is developed with React and bundled using Vite, ensuring a fast, smooth user experience with real-time interactions and clean state management.

At the heart of our emission tracking engine is Google Gemini, which analyzes both image and text inputs from receipts or user logs. These raw insights are passed through LangChain, which acts as a retrieval-augmented generation (RAG) layer. LangChain cross-references Gemini’s extracted content with a custom Carbon Item Dataset. The data set served as a structured knowledge base containing emission data for various consumer goods.

To provide contextual accuracy, the carbon dataset is embedded and indexed, enabling LangChain to retrieve relevant emission factors in real-time. The backend, built with Flask, orchestrates this workflow, handling user inputs, communicating with Gemini and LangChain, storing results in a SQLite database, and serving API responses to the frontend.

The platform is designed to be scalable and community-driven, with features like leaderboards, EcoCoins, and partner rewards managed within the Flask backend and synced to the UI. Every component from image analysis to gamified feedback works together to help users better understand their carbon footprint and take actionable steps toward sustainability.

Challenges we ran into

One of the key challenges we faced was connecting our React frontend to the Flask backend in a way that maintained responsiveness while handling real-time data requests from the AI pipeline. This required careful handling of CORS policies, API structure, and asynchronous data flow.

Another technical hurdle involved updating the SQLite database with dynamic user data, especially tracking and storing EcoCoins as users performed actions. Designing a lightweight yet scalable schema that could track emissions, coins, and leaderboard data took thoughtful iteration.

We also encountered challenges with prompt engineering for Google Gemini. Getting accurate carbon estimates from diverse receipts meant refining the input prompts and iteratively adjusting them to produce reliable output, especially when dealing with edge cases or vague product names.

Lastly, we spent time on fine-tuning the interaction between Gemini and LangChain, ensuring that the model could retrieve and synthesize emission data effectively. Balancing performance, accuracy, and interpretability across the AI layer was one of the more timeconsuming parts of our build.

Accomplishments that we're proud of

We are incredibly proud of developing a tool that can help individuals track their impact on the environment and make the world greener one step at a time. Successfully integrating complex AI technologies to create a user-friendly platform that can genuinely help users reduce carbon emissions and learn more about the impact they are making in their day-to-day lives is an achievement that resonates with our core values. Additionally, our ability to work as a cohesive team, leveraging each member's strengths to overcome challenges, stands as a testament to our dedication and collaborative spirit.

What we learned

Throughout this journey, we’ve gained invaluable insights into how AI can empower individuals to make more sustainable choices in their everyday lives. We explored the intricacies of receipt scanning, image-to-text processing, and contextualizing consumer data to estimate carbon emissions with accuracy. This project deepened our understanding of prompt engineering, multimodal AI capabilities, and the challenge of translating raw purchase data into meaningful environmental insights. Most importantly, we learned the importance of designing with intention and empathy, creating a tool that is not only technically impressive but also intuitive, rewarding, and impactful. This experience reinforced the power of collaboration, creativity, and purpose-driven innovation in addressing real-world problems.

What's next for ReceiptReveal

We plan to introduce a competitive leaderboard, allowing users to challenge friends and communities to reduce emissions together. In parallel, we aim to open the platform to organizations and institutions, giving them tools to track collective carbon reductions. These efforts could be tied to incentives like lower interest rates from eco-conscious financial institutions, encouraging climate action at scale.

To enhance the accuracy of our emission estimates, we intend to expand our carbon item dataset, enabling more precise contextual analysis across diverse purchases. Another exciting direction is building a network of partner stores and sustainable brands, where users can redeem EcoCoins for exclusive discounts. This not only makes the platform more engaging for users but also introduces a sustainable revenue stream, aligning profitability with planetary impact.

Long-term, we envision ReceiptReveal becoming the go-to personal climate companion a smart, community-powered ecosystem that drives meaningful environmental change through daily decisions.