7 Degrees | Devpost

Logo
Landing Page
Home Page
Connections Page
QR Authentication Page
Tools We Used

The Problem 💭

Conversations are how relationships start, but they’re fragile. After a hackathon or networking event, you might have talked to 20 people, but a week later, you can’t remember who mentioned what. Was it Sarah who said she was struggling to break into machine learning? Or Toby who’s in the entrepreneurship club? Those are the details that make follow-ups meaningful, but they’re also the first to fade.

The friction comes in two forms: The Human Problem: We connect best when we find common ground shared struggles, shared passions, or ambitions. But without a system to surface those overlaps, most of it gets lost. Follow-ups become shallow (if they ever happen), and opportunities for real connection never happen. To add to all of that, managing and enhancing your relationship with people takes a lot of effort, so can technology solve that problem?

The Technical Gap: Current tools still require human effort and they don't leverage the best technologies for the purpose of relationships. We thought of it from the ground up and realized that storing relationships in a knowledge graph that is powered by AI (chat with it, go on a phone call with it) alligns perfectly with how human beings think. So in some way it's a 2nd brain, a memory vault for who you care about or want to grow your relationship with

Inspiration 🌎

Recently, there has been a boom in virtual meeting bots that can summarize and take actions on meeting notes. One popular e.g. is granola.ai , but there are MANY more, but what about face-to-face interaction?

The name 7 Degrees comes from the “seven degrees of separation” idea, which suggests that any two people in the world are connected through just a handful of relationships. Our project extends this concept by focusing on professional and personal networking, helping people map and strengthen those connections through meaningful conversations.

What it does 🤔

7 Degrees is a relationship intelligence platform that captures, organizes, and enhances conversations. Here’s how it works:

🎙️ AI Audio-to-Text + Analyzer + Summarizer: Conversations are transcribed in real time and analyzed for key points, giving users a structured summary of who said what.
🧩 Graph Database for Connections: We store conversations in a graph structure that maps relationships, making it easy to see how people, topics, and ideas are linked.
🎨 Sleek User Interface: Our web app offers a clean, intuitive dashboard where users can view their conversations, relationship graphs, and insights at a glance using Next.js and TailwindCSS.
🔐 Authentication & QR Security: Conversations and connections are secured with authentication flows and QR-code-based meeting check-ins, ensuring that only verified interactions are stored. recordings won't be saved unless you get your QR code scanned by the other person. The other advantage of that is that the user will automatically get connected and the meeting notes and knowledge graph will be shared with them

How we built it 🛠️

Frontend: Next.js with TypeScript, TailwindCSS, and shadcn/ui for a smooth and modern UI.
Backend: Flask + Python, running endpoints for transcription, conversation analysis, and summarization. We integrated OpenAI Whisper for transcription and Gemini for extracting key facts and summaries.
Database: Supabase (Postgres + auth), for storing users and conversations, with pgvector for embeddings. Connections between users are modeled like a graph to reflect the "7 Degrees" concept.
Infra & Dev Tools: Leveraged Windsurf IDE to combine and integrate our code smoothly, deployed on Vercel, and used Postman for endpoint testing.

Challenges we ran into ⚠️

LLM output consistency: Ensuring our AI analysis always returned proper relationships between users Frontend-backend sync: Coordinating Next.js with Flask endpoints and debugging CORS issues.
🗂️ Database design with context engineering in mind: Deciding how to represent relationships as a graph inside Supabase while still making it queryable for summaries and follow-ups.

Accomplishments that we're proud of 🏆

Built a platform that anyone can use to effortlessly capture, analyze, and summarize conversations, making networking and building relationships way easier.
Made our platform accessible and intuitive, so users can get insights without a steep learning curve.
managing to connect all the itegrations that we had in mind (like What'sApp)

What we learned 🧠

• Graph databases paired with AI can really help visualize and understand relationships.
• How much an AI-assisted IDE like Windsurf can speed up collaborative development.
• Making a platform that’s both easy to use and secure takes careful balance, but it’s worth it.

What's next for 7 Degrees ➡️❓

Deeper integrations: Connect with LinkedIn, Slack, or email so you can see past context before every interaction.
Personalized AI memory: Smarter reminders and suggested follow-ups based on your conversation history.
Analytics Dashboard: Insights on network growth and relationship depth over time.