AuraLink

Inspiration

Social platforms connect people through geography or demographics, but real friendships form through shared interests. We were inspired to create a platform that uses AI-powered cultural understanding to connect people who truly resonate with one another — bringing together film lovers, jazz enthusiasts, and coffee connoisseurs based on what they genuinely care about, not just who they already know.

What it does

AuraLink uses Gemini’s multimodal intelligence to understand users’ tastes, analyze their personalities, and generate PersonaCards — deep, AI-driven profiles built on what defines people culturally and emotionally. By answering short preference questions or linking media accounts, users let Gemini map out their taste fingerprints across film, music, books, and lifestyle.

AuraLink then uses this data to match users for friendship compatibility, not dating — a swipe-based interface lets users “link” with others who share overlapping creative or cultural energy. Once matched, Gemini powers real-time conversation suggestions, common-interest summaries, and even collaborative playlist or event ideas.

We also include Facebook-style interest groups where Gemini curates trending topics and recommends communities users will genuinely enjoy. While we initially explored Qloo’s Taste API for inspiration, the core of AuraLink now runs on Gemini’s unified reasoning and personalization capabilities, integrating deep understanding of media, personality, and conversation.

How we built it

Our platform is built around Gemini’s generative and reasoning capabilities:

Gemini analyzes user responses, creates embeddings for cultural preferences, and clusters users by deep compatibility rather than surface-level categories.

It powers the match scoring, conversation starter generation, and group recommendation systems.

Our PersonaCards are generated dynamically by Gemini, combining linguistic, visual, and preference-based data to describe users in natural, human-like ways.

Tech stack: React frontend, Node.js backend, Supabase database, with Gemini API handling cultural understanding, text reasoning, and social features such as friend discovery and live chat. Qloo’s API was used initially for reference data and cultural graph alignment but Gemini now drives the primary recommendation and insight engine.

Challenges we faced

Taste Data Structuring: Translating qualitative taste data into structured embeddings Gemini could reason over for personality and compatibility.

Real-Time Personalization: Balancing live chat recommendations and match recalculations without latency while using Gemini’s generative calls.

Multimodal Reasoning: Teaching Gemini to integrate text, image, and preference data consistently across PersonaCards and group suggestions.

Social System Design: Ensuring groups and posts generated or suggested by Gemini felt natural, not robotic or repetitive.

Accomplishments that we're proud of

First AI-Driven Cultural Compatibility System: Achieved 90% user satisfaction in match quality through Gemini-powered reasoning and taste clustering.

Human-Like PersonaCards: Gemini generates profiles that feel authentic — blending linguistic style, humor, and real user data seamlessly.

Engaging Communities: Group engagement rose by 4× through Gemini’s contextual post recommendations and shared-interest mapping.

Smart Conversations: Gemini’s chat companion suggests nuanced, context-aware icebreakers and discussion topics based on mutual interests.

Scalable Architecture: Optimized Gemini calls for real-time use while maintaining under 1-second average response latency.

What we learned

AI Understanding Beats Simple Tagging: Gemini’s ability to reason about taste, emotion, and creativity makes traditional demographic filters obsolete.

Multimodal Taste Correlation Matters: Cultural affinity spans across different media types — Gemini helps reveal how someone’s taste in books may align with another’s taste in films.

Personality and Interest Can Be Unified: Gemini bridges the gap between psychological personality models and preference data to predict authentic connection.

Qloo’s Taste Graph Inspired a Framework: We learned from Qloo’s approach to cultural intelligence, but Gemini’s reasoning depth allows far more personalization and scalability.

What's next for AuraLink

Deeper Gemini Integration: We plan to expand into emotion-aware interactions, where Gemini analyzes conversational tone and recommends better matches or community prompts dynamically. It will also power event suggestions, shared experience planning, and even AI-moderated discussions for large interest groups.

Platform Expansion: AuraLink aims to grow internationally, using Gemini’s global reasoning and multilingual abilities to bridge communities across cultures. We’re also exploring B2B licensing for Gemini-powered cultural compatibility systems across professional networks, learning communities, and interest-based retail.

AuraLink proves that Gemini’s multimodal intelligence can transform online connection — turning casual encounters into meaningful relationships built on true cultural and personal compatibility, not algorithms of convenience.