Inspiration
Valores came from my own experience with apps like Tinder and Yubo. I used them, and I noticed that I was swiping mostly on looks, not on how someone actually thinks or behaves. Your first impression of a person is based on their photo, which isn’t really fair to either side.
I wanted a place for people who would rather talk about how they think, what they feel and who they are than how they look. The name comes from Latin and Spanish for “values”, and the tagline says the rest: More than your looks.
What it does
Valores is a simple mobile app for meeting people who think similarly to you. The twist is that there are no photos at all.
Instead, every user builds a board: a personal map of their mind made of blocks like values, travel and hobbies, or custom blocks with their own names and notes. Other people discover you by swiping through boards. If you both swipe right, you match and can start a conversation.
To be interesting here, you have to say something interesting about yourself. You don’t show your face, you show your mind.
How I built it
I planned the architecture and wrote the whole backend myself, in Python with FastAPI. It’s split into routers, services and models, with separate modules for auth, boards, discovery, swipes, likes, conversations and messages. Data and authentication live in Supabase (PostgreSQL), with tables for user profiles, boards, swipes, likes, conversations and messages. The backend is deployed on Render. Chat works by storing messages in the database and fetching them from the app.
The frontend is a Flutter app, and I’m honest about this part: AI wrote it. I’m a backend developer and barely know Flutter, so I described what I needed and let AI build the screens (login, swiping, board editor) on top of my endpoints. I mostly used gpt-oss:120b through the Ollama cloud API connected to the Zed Agent Panel, with occasional help from other models, including Claude Sonnet.
The whole thing took about a week of real work: roughly 3 intense days of building, and then around 4 days of deployment and getting it onto a real phone.
Challenges I ran into
Designing the board. This was the hardest problem. I originally imagined a free, draggable canvas, like the node editors in n8n. The AI built something much simpler: you add blocks and they arrange themselves. It turned out to be more comfortable to use, so I kept it. It also leaves room to add new ways of building and decorating boards later.
Working with a frontend I didn’t write. I couldn’t rely on my own Flutter knowledge, so the quality of the result depended on how well I explained what I wanted. Sometimes the AI misunderstood my idea and built something different.
Cutting scope. Valores was meant to be partly a small social network, where people share thoughts, books and music. I cut it to finish the core idea in time.
Free-tier hosting. The backend runs on Render’s free tier, so the first request after a period of inactivity can be slow while the server wakes up.
Accomplishments that I’m proud of
I built and deployed a complete working app: accounts, board editor, swiping, matching and chat, installable on a real phone. The backend is one of the most serious projects I’ve done, with a clean structure and a real database and authentication. The idea itself. I honestly think I’d use it myself, which isn’t something I can say about every project. I made a promo video and can show a live demo. What I learned Practical Flutter: how to change the app icon, and how to test on a real phone while my computer acts as the server on the same network. Working with AI as a builder: it can get you into a field you don’t know, but the result depends on how clearly you describe what you want. Security: my API currently takes the access token as a request parameter instead of a Bearer token in the header. It works, but it isn’t the professional way, and I know what to fix first. Simpler can be better: the board I ended up with is not what I first imagined, and it’s the better version.
What’s next for Valores
Valores is a portfolio project that I’ll keep developing:
Bring back the social side: a space where people share thoughts, books and music, so Valores becomes more than a matching app. Richer boards: more block types and more ways to decorate and customize a board. User blocking and other safety features. Finding the first community: I plan to start with short videos about the app on Instagram and test it with real users.
Log in or sign up for Devpost to join the conversation.