Dear Diary, take me there

Dear Diary, Take Me Here

Travel recommendations today are usually based on where you've been or what you've searched for. We wanted to explore a different idea. What if travel recommendations started with how you feel instead?

Dear Diary, Take Me Here is an AI-powered journaling experience that transforms a personal journal entry into a meaningful travel recommendation. Instead of filtering destinations by popularity or price, our app analyzes the emotions and themes in your writing, reflects them back to you, and suggests a destination that matches your current mindset. It then recommends real accommodations so you can turn that inspiration into an actual trip.

Inspiration

Many of us journal to process our thoughts, but those reflections rarely lead to action. We were inspired by the idea that travel can be restorative, grounding, and deeply personal.

Instead of asking users, "Where do you want to go?", we wanted to ask, "What do you need right now?"

By combining AI with travel discovery, we wanted to create something that feels more thoughtful than a traditional booking platform. The journey begins with self-reflection instead of a search bar.

How We Built It

We built the application using:

React + TypeScript for the frontend FastAPI (Python) for the backend Large Language Models (LLMs) to analyze journal entries and identify emotional themes Stay22 API to retrieve real accommodation recommendations Vercel for deployment

On the frontend, we used an adapter pattern to separate the UI from the backend response format. This made it easy to develop with mock data while keeping the transition to the live API seamless.

Challenges We Ran Into

One of our biggest challenges was designing AI interactions that felt supportive without sounding robotic or overly clinical. We wanted the reflection to feel like someone genuinely understood your writing, not like it was diagnosing you.

Another challenge was developing the frontend and backend in parallel. We solved this by agreeing on a shared API contract and using an adapter layer so the frontend could remain stable while the backend was still being built.

We also had to integrate AI-generated insights with live accommodation data, while gracefully handling cases where information like ratings or hotel availability might be missing.

What We Learned

This project taught us that great AI experiences are about much more than the model itself. The way information is presented, the wording of the reflection, the pacing of the experience, and the overall user interface all contribute to building trust.

We also gained valuable experience integrating multiple APIs, designing clear frontend and backend contracts, and building an application where AI enhances the user experience instead of replacing it.

What's Next

We'd love to expand the experience by generating personalized itineraries, recommending local experiences based on the user's emotional goals, and helping users revisit past journal entries to see how their travel preferences evolve over time.

Our goal is to encourage people to see travel as more than just a destination. Sometimes, the best place to go starts with understanding where you are emotionally first.