-
-
User Flow Diagram
-
-
Inspiration
Every year, millions of animals suffer due to cosmetic testing — despite the growing global awareness of cruelty-free ethics. Yet, consumers often face greenwashing and misinformation that make it hard to verify brand claims. We wanted to empower conscious shoppers to make ethical decisions without spending hours cross-checking certifications, parent companies, or hidden animal testing policies.
ConsciousCart was born from the idea that data science and agentic AI can simplify ethical consumerism and make cruelty-free beauty accessible to everyone, not just privileged consumers.
What it does
ConsciousCart is an AI-powered verification agent that autonomously determines whether a cosmetic brand is cruelty-free.
It:
Checks verified databases and certifications (like Leaping Bunny and PETA). Identifies parent company relationships that may test on animals. Performs intelligent web searches when needed. Returns results with a confidence score, ethical reasoning, and cruelty-free alternatives. Learns and updates over time with caching and a memory buffer to improve accuracy and efficiency.
Users simply enter a brand name and instantly see whether it’s cruelty-free, why, and what ethical alternatives exist.
How we built it
Language: Python 3.10 UI: Streamlit (for interactive brand verification) Database: SQLite for persistent brand and certification data Agent Framework: Claude-powered autonomous reasoning agent Search System: Mock web search integrated (production-ready for MCP Brave Search) Data Science Layer:
Multi-source synthesis Knowledge graph-style reasoning Intelligent caching Confidence scoring system
We structured the system so that known brands are verified in 1–2 tool calls, while unknown brands may trigger 5+ tool calls for deeper cross-verification.
Challenges we ran into
Parsing and cleaning fragmented data from certification databases. Preventing false positives from brands that claim cruelty-free status but have parent companies that test. Designing a realistic multi-tool agent workflow that can balance speed and reliability. Building a responsive UI that explains reasoning clearly without overwhelming users. Managing local and mock web data sources efficiently within the demo time limit.
Accomplishments that we're proud of
Built a fully working AI agent that can autonomously verify cosmetic brands. Integrated tool use and caching for faster reasoning over time. Designed a clean, interactive Streamlit interface with transparent explanations. Created a foundation for scalable, ethical AI consumer tools. Developed a framework that can easily expand to other product categories (e.g., fashion, household, or food).
What we learned
How to design autonomous multi-tool AI agents for real-world use cases. How to connect data science and ethical reasoning for consumer empowerment. Importance of trust, explainability, and transparency in AI systems. How to structure multi-source data pipelines (databases, APIs, and web results). Learned a ton about ethical data handling and sustainable product design thinking.
What's next for ConsciousCart
Integrate MCP Brave Search for live web verification. Connect to official certification APIs (PETA, Leaping Bunny). Expand to 100+ verified brands and user profiling across sessions. Add personalized recommendations based on budget and values. Build a Chrome extension so users can verify products directly while shopping online. Extend the system to fashion and household goods for broader ethical consumer coverage.
Log in or sign up for Devpost to join the conversation.