Inspiration The idea for Gift Smart came from wanting to create the perfect gift for my girlfriend. I realized how stressful gift-giving can be—especially when you want it to be meaningful, thoughtful, and within budget. I wanted a way to take the guesswork out of gifting and help people feel confident in what they give. That personal need sparked the idea for a smart, AI-powered tool that helps users generate curated gift ideas based on age, interests, and budget.
What it does Gift Smart is a full-stack web app that helps users generate thoughtful gift ideas tailored to the recipient. After entering details like the person’s name, age, hobbies, and your relationship to them, the app uses AI to generate a list of gift suggestions. Each gift includes a name, price, description, and link, and may also include DIY instructions or event-based gifts like shared experiences. The goal is to provide a variety of thoughtful, personalized ideas that suit different personalities and budgets.
How we built it Frontend: Built in React using Supabase Auth for login and state management for giftee data.
Backend: FastAPI serves as the API layer, handling prompts and responses from the AI.
AI Integration: Initially used Gemini API to generate gift suggestions from prompts. Later integrated Perplexity AI for flexible LLM handling.
Database: Supabase (PostgreSQL) for storing user data and giftee profiles.
Deployment: Frontend is hosted on Vercel, and the backend runs on Render.
Challenges we ran into Parsing AI responses: Ensuring the AI output was valid JSON was tricky, especially with multiline objects and edge cases.
Prompt engineering: Crafting a clear and specific enough prompt to return useful, structured gifts took lots of trial and error.
Cross-origin requests: Dealing with CORS and environment variables when connecting the backend and frontend on different hosts was another challenge.
Data structure consistency: Building a schema that could support multiple gift types (DIY, event-based, etc.) in a unified format required thoughtful planning.
Accomplishments that we're proud of Successfully implemented AI-generated gift ideas with minimal user input.
Created a modular backend that can swap out LLMs (e.g., Gemini ↔️ Perplexity) without major refactors.
Designed a user-friendly interface that guides users through creating giftee profiles and browsing suggestions.
Developed DIY and event gift support with nested structured fields like steps, materials, and event details.
What we learned How to prompt large language models effectively for structured, JSON-based responses.
The importance of validating and cleaning AI-generated data before using it in production.
How to architect a full-stack app with clean separation between frontend, backend, and AI services.
Practical skills in integrating Supabase, FastAPI, and modern AI tools in a real-world project.
What's next for Gift Smart Add AI-generated images for each gift suggestion to make the cards visually engaging.
Let users save favorite gifts to a wishlist and mark gifts as "purchased."
Add reminders for upcoming birthdays and holidays.
Log in or sign up for Devpost to join the conversation.