Hygia Lens

Hygia Landing Page
Opening camera through the bottom menu
Scanner UI opens up
Our AI powered backend shows what's used in the shown product and safety rating of these ingredients
User scans up a supplement label

Inspiration

Finding out about all the ingredients used in the products that we consume on a daily basis is a hard tasks, foods usually have n number of ingredients on their label and searching each of them manually while keeping track of how safe they are is a tedious process. We wanted to simplify this for users of all levels, whether they are tech savy or not. Anyone who has an internet access and a smartphone or device with a camera they shall be able to use Hygia Lens.

What it does

Hygia Lens let's you scan ingredients label of any food, products, supplements or cosmetics that you have and it tells you what all ingredients are used and how safe they are.

We rank the safety of these items from 0 to 10, where 0 is the most safest and 10 being the most hazardous.

How we built it

We have currently built an early prototype for our application, where user can scan the ingredients label and view the information easily. We used Next.JS for the frontend and TailwindCSS for styling, and we are using Node & ExpressJS with Typescript on the Backend. We are also using Google Images API & Open AI for handling the image-to-text & ingredients safety assessment. We have deployed frontend on Netlify platform and backend is deployed to our own VPS. We also tried creating different parts of UI with bolt.new. And we are using supabase for implementing the authentication part on the frontend.

Challenges we ran into

Testing out which LLM model to go with was a bit tricky part, and we did trial and error with bunch of these mainstream LLM models.

Accomplishments that we're proud of

Went from Idea to POC to this MVP in around 2 weeks time We also learnt alot on the way, things like streaming server response and using supabase.

What we learned

We learned about using Google text to image api for extracting text out of images. We also learnt how llms like chatgpt stream their response and how it's shown in the UI.

What's next for Hygia Lens

We are planning to launch Hygia Lens as a mobile application, with clean and minimalistic UI, and we will also introduce features like LLM Chat with our Health Agents and integrate more database features like saving user scans and history of scans etc and for that we will be using supabase.