WeThinkWeConnect

Do you have half-baked thoughts or unsettled disputes? WeThinkWe Connect is the go to website for you. An anti-social network from having chain of thoughts to cold wars. We have got you covered.

Comment

Inspiration

People are interactive on a lot of platform but they don't know what kind of interaction boosts their personality or what is the sentiment of their thoughts and beliefs. They don't know if they are interactive in a right way or not as they randomly spam dm and comment of multiple users. To negate this I came up with WeThinkWeConnect.

What it does

WeThinkWeConnect is an "anti-social" network designed to prioritize depth over dopamine. The app removes traditional metrics like "likes" and "followers" to focus on three pillars: vulnerability, radical empathy, and structured conflict.

It features three distinct modules:

Unfinished: A feed where users post raw, incomplete thoughts. Instead of likes, our Gemini AI analyzes the sentiment (Melancholic, Whimsical, Dark) and assigns a mood color. Users can only interact by "extending" the thought, not judging it.

EchoSwap: A space where you cannot reply to a belief until you have successfully rewritten it from the other person's perspective. It forces you to understand before you are understood.

Conflict: A real-time debate lobby. Two users enter with a disagreement, but the chat is locked. To unlock the ability to speak, both users must move a "Convergence Slider" closer together, physically symbolizing compromise before words are exchanged.

NOTE: The whole App in moderated by AI

How I built it

The application is built on a modern, real-time stack:

Frontend: React.js with Tailwind CSS for a glassmorphism aesthetic.

Backend: Node.js and Express.

AI Integration: I used Google gemini-3-flash model for real-time features (like the "Spark" auto-complete) and Gemini-3-pro model for deeper sentiment analysis and moderation.

Real-Time Communication: Socket.io handles the live synchronization in the 'Conflict' feature, ensuring both users see slider movements instantly.

Database: Neon (Serverless PostgreSQL) stores user profiles and thoughts chaining them together.

Challenges I ran into

  1. The AI features (Sentiment Analysis and Spark) initially slowed down the UX. I solved this by implementing the lightweight Gemini Flash model for instant UI suggestions and the heavier Pro model for background analysis.

  2. Synchronizing the "Convergence Slider" in the Conflict Room between two users without lag was tricky. I had to optimize our Socket.io events to prevent race conditions where one user unlocks the chat while the other is still locked.

  3. Deploying was the real issue and a hassle since it was my first time deploying a monorepo created in react + vite and node + express, it took me almost an hour to figure out things.

Accomplishments that I am proud of

The Spark Feature: I successfully created a custom React hook (useGeminiSpark) that helps users overcome writer's block by suggesting 3 distinct, poetic continuations of their sentences in real-time.

Gamified Emotional Profile: Instead of a generic profile, I built a Dashboard that tracks Vulnerability, Empathy, and Diplomat scores, giving users a visual representation of their emotional footprint.

No-Like Architecture: I successfully built an engaging social feed that retains users without a single "Like" button, proving that content can drive engagement through curiosity rather than validation.

What I learned

I learned how to deploy a monorepo made with react + vite and node + express successfully. I deployed the backend server on Render and frontend on Vercel.

I learned that the difference between a generic AI response and a "human-sounding" one lies in strict prompt constraints.

What's next for WeThinkWeConnect

Creating welcome Email sending logic in the backend. Integration of OAuth for Google login/signup. Integration of Stripe payment gateway. Advertising it to the world.

Share this project:

Updates