SnapSage

Ask It. See It. Know It.

Comment

Inspiration

I wanted to make information visual, fast, and easy to digest. Instead of reading long paragraphs or be taken to youtube for a visual example, what if you could see an answer instantly? SnapSage was born to combine the power of AI and visual search to make learning and discovering feel natural, quick, and inspiring.

What it does

SnapSage takes any question you ask, generates a short, clear description, and finds relevant images that explain the topic visually. In just seconds, users get instant understanding through a mix of words and pictures . It is perfect for studying, exploring ideas, planning trips, or satisfying curiosity.

How we built it

We built SnapSage using:

  • FastAPI for the backend server
  • Google's Gemini AI to generate short answers and smart keywords
  • Google Custom Search API to retrieve the best matching images
  • Python for processing and error handling
  • CORS middleware to allow frontend integrations easily
  • Html for UI building
  • JavaScript

We focused on clean async handling, relevance checking, and maximum image quality to give users the best experience.

Challenges I ran into

Ensuring image relevance: Not every search result matched the question. I had to build an extra AI step to check titles/snippets.

Limited Knowledge: I am a first year Science student with almost no experience in programming. I was literally learning as I was working.

Host Connection: It was rather difficult connecting our backend (Python) to our frontend (Html and JavaScript) without the use of a local host.

Accomplishments that I'm proud of

Creating a fully working prototype that answers a question and gives visual + text explanations.

Building a double-layer AI system (answer generation + relevance checking) that makes results smarter and more reliable.

Keeping the experience lightweight, fast, and extensible for future versions.

What I learned

When I started SnapSage, I had very little experience with the technologies we ended up using. Throughout this project, I learned everything from setting up servers with FastAPI, handling APIs like Gemini and Google Search, managing async operations, securing environment variables, to building a complete end-to-end system.

Every challenge pushed me to research, adapt, and problem-solve in real time. This project was a full crash course in backend development, AI integration, and real-world app design and I'm proud of how much I grew.

What's next for SnapSage?

Snap Collections: Let users save topics and build mini-boards for study, travel, or idea inspiration.

Voice Search: Allow users to ask questions by speaking, not just typing.

Offline Packs: Download visual info packs on popular topics (e.g., “Space Exploration”, “Ancient Wonders”).

Smart Explore Mode: Suggest related questions and topics automatically after a search.

Mobile App: Expand SnapSage into a mobile app to make instant learning even more accessible.

Share this project:

Updates