SOCIUS

💡 Inspiration: Never Lose a Valuable Connection In today's fast-paced professional world, networking is your greatest asset. Yet, how often do you exchange contact info at a bustling event, only for that vital interaction to become a forgotten business card or a missed LinkedIn connection?

Our inspiration is to solve this "professional memory gap." We've built an AI agent system that acts as your personal notification hub, ensuring you never overlook a meaningful interaction. It moves beyond simple contact exchange, offering a tool for remembering, synthesizing, and acting on the connections that truly matter.

✨ What it does: Seamless, AI-Driven Contact Exchange and Synthesis SOCIUS (our agent application) eliminates the awkward, manual process of exchanging contact information. Forget QR codes or typing names.

Instead, a simple iMessage ping from one person to our core agent initiates a sophisticated, background handshaking process:

Your phone pings the SOCIUS Agent on your end.

Your agent securely communicates with the other person’s agent.

The agents pass data to the core node, which then processes the interaction.

Crucially, the system uses AI to summarize the interaction, highlighting key discussion points and any professional commendations or action items, which are then delivered as a LinkedIn notification or summary.

It doesn't just deliver contact info; it delivers context and follow-up potential.

🛠 How we built it: A Powerful, Multi-Stack AI System We engineered SOCIUS using a modern, robust, and scalable technology stack designed for cross-platform performance:

Core Logic & AI: Python (Backend), LangChain (Orchestration), and Anthropic - Sonnet Haiku 4.5 (for sophisticated interaction summarization and commendation extraction).

Data & Speed: Redis (High-speed Caching) and Sanity (CMS) (Flexible Content Management for evolving features).

Client & Mobile: Swift (iOS Integration for the seamless iMessage trigger) and TypeScript (Robust Web/App Components).

Testing: Postman (Ensuring rigorous, reliable End-to-End communication between agents).

Infrastructure: Lookcloud (Cloud Service Integration).

⛰ Challenges we ran into: Focus and Integration Our biggest hurdle involved managing the complex network topology of agent-to-agent communication and ensuring secure data exchange. This led to challenging trade-offs, including an initial pivot from a direct-to-LinkedIn exchange (due to API constraints) to a more reliable, agent-mediated notification structure that ultimately allows for richer, summarized data delivery.

🌟 Accomplishments that we're proud of: Agent-to-Agent Reliability We are immensely proud that the entire agent-to-agent communication pipeline works reliably. We successfully built a system where a simple iMessage trigger can initiate a complex, multi-stage AI process, retrieve data, summarize it, and deliver a valuable, actionable notification.

🧠 What we learned: Full-Stack Systems Design We gained hands-on expertise in building a truly full-stack, distributed system—mastering everything from low-level mobile integration (Swift) to high-level AI orchestration (LangChain), and ensuring state consistency (Redis).

🚀 What's next for SOCIUS: Scaling the Connection Hub Our immediate next step is making SOCIUS production ready—focusing on user experience refinement, security audits, and preparing for public release. Beyond that, we plan to integrate actionable follow-up suggestions based on the summarized commendations, turning SOCIUS from a notification tool into a powerful relationship management assistant.