Date Yelper

An AI-powered date planner that creates personalized date plans based on your interests, budget, location, and schedule—so you can stop planning and focus on the date itself.

Comment

Inspiration

Many people nowadays are bored of their relationships, including a friend of mine. She told me that she felt like she and her partner weren't making an effort for each other anymore, and that they didn't go on any special dates anymore. Online, many people have talked about the same thing - why did we stop making an effort? Planning fun dates is on a downtrend. I get it - work and everything else just takes so much of your time and energy, that the last thing you want to do is research and plan an entire date. So Date Yelper will do it for you!

What it does

Date Yelper will plan an entire date for you based on your requirements - location, budget, interests, time, reviews, etc. Instead of searching up so many websites and trying to find new, fun things to do, Date Yelper will do it for you! You can directly chat with our date agent Nate, give him feedback and he will tailor the entire itinerary to your needs!

You can use this account login for demo purposes: Email: 1234demo@gmail.com Password: 1234demo

How I built it

Honestly, in the past, I tried making my own web apps, but it always got so messy and uncoordinated, I tried a different approach this time. This time, I tried an actual, organized workflow which started off with a general PRD (Product Requirements Document). From that, I generated my requirements for my MVP. I fed my PRD and MVP requirements to my LLM and it generated a coding workflow-specific plan that followed vertical integration of steps. For Date Yelper, all of the steps from idea to MVP consisted of 5 phases -

  1. Frontend
  2. Supabase data & auth integration
  3. Yelp API integration
  4. Agent Core and integration with Yelp API data input
  5. Connection with UI and cleanup

Challenges I ran into

Planning everything was the most challenging part. Getting clarity on the exact features I wanted for the MVP and for the more fleshed out product was tricky, even though I bounced off ideas with my LLM. Doing this workflow and vertical integration was tough to navigate at first, because I wasn't quite sure on what was right or wrong or when I could trust the LLMs judgment on which path to go. It was especially confusing differentiating the PRD and MVP requirements, and getting everything I needed in the PRD.

Accomplishments that I'm proud of

I am not a fullstack developer, and this is one of my first fleshed-out projects, so I'm really proud of that! It felt really good to finally have a structured workflow that I can continue to use in the future for other projects. I think that has been my bottleneck for all of the projects I've done in the past so far.

What I learned

Even though the AI is capable and knows of thousands of ways to do things, I learned how to judge whether their decision or recommendation was fitting or not. At the end of the day, you have to know how to properly communicate with the AI, and have to get clarity on exactly what you want. Without that roadmap or north star, you'll keep wandering, not reaching where you want to go.

What's next for Date Yelper

User feedback loops; more detailed and higher quality filtering of activity recommendations; weather updates; act as platform to share good date places where people can share with each other; recurring learning engine for better dates itinerary based on feedback.

Built With

Share this project:

Updates