Demo Video (Youtube):

Devpost: https://devpost.com/software/labellens

Ingredients (especially in skincare, cosmetics, and medicine!) are overwhelming. With thousands of obscure ingredients and conflicting advice, users need a neutral, science-backed tool to decode ingredient labels. LabelLens was born to democratize ingredient literacy!

LabelLens empowers its users by allowing for simple, un-biased, and convenient way to look up ingredients. Users can either take a picture of an ingredient list to get a quick summary of each ingredient, look up a specific product's information and ingredients, or look up a specific ingredient themselves.

Most competitors fail users in two ways:

• Bias for Profit: Many apps quietly partner with brands, skewing results or hiding unfavorable data.
• Stale Data: Manual updates can’t keep pace with new ingredients (e.g., vegan alternatives, novel actives).

LabelLens combats both with:

• RAG + Google Gemini: Our AI cross-references real-time research and vetted databases, avoiding human curation bias.
• Zero Brand Partnerships: We prioritize user trust over monetization, ensuring analyses are never influenced by third parties.

• React.js with Tailwind for the frontend
• Node.js for the backend
• Google Gemini and GCP (Docker, Kubernetes) for the AI, Weaviate for the vector database
• Data was collected from various reputable sources, such as the FDA.

1. Clone the repository

git clone https://github.com/Deeppcodes/LabelLens.git

2. Navigate to the project directory

cd LabelLens

3. Create a .env file with your API key

VITE_GOOGLE_API_KEY=<YOUR_KEY_HERE>

4. Install dependencies

npm install

5. Start the development server

npm run dev

• We're proud to have learned to combine custom RAG algorithms with Google Gemini

• Comparing two items (e.g. If two items are )
• Providing alternatives (e.g. I like a product, but it contains an ingredient I do not like, is there an alternative of the same level(?) but without the ingredient?)