LinkForge

LinkForge is an AI-powered decentralized networking platform designed to enhance the experience of attending blockchain and tech events. It transforms traditional networking into a gamified, engaging, and rewarding process by helping attendees connect meaningfully, complete personalized challenges (quests), and earn experience points (XP)—all tracked transparently on-chain.

At its core, LinkForge uses AI agents to analyze user interests, generate networking suggestions, and craft quests tailored to each participant’s profile and event theme. These quests might involve meeting specific attendees, visiting booths, attending sessions, or completing skill-building tasks—fostering organic and valuable interactions.

All interactions, including connections and quest completions, are recorded via smart contracts, and users earn XP for every action. This XP contributes to a real-time leaderboard that adds a fun, competitive layer to the event. The more you engage, the more XP you earn—and top performers get featured!

Event organizers can create and customize events through the platform, defining interest categories, branding, and timelines. Attendees simply connect their wallet using Privy, set their interests, and are ready to network with purpose.

Multi-chain support ensures broader reach and compatibility, currently deployed on Celo, Base Sepolia, and Rootstock Testnet.

Whether you're a developer, builder, investor, or enthusiast, LinkForge makes event networking smarter, more rewarding, and future-ready with Web3 and AI.

🛠 How It’s Made LinkForge was built with a modern tech stack combining Web3 infrastructure, AI, and full-stack web development tools to deliver a seamless and intelligent networking experience.

🧩 Frontend We used Next.js to build a fast, scalable, and responsive frontend. The framework allowed us to implement server-side rendering and API routes efficiently, making the user experience smooth across devices. Tailwind CSS and ShadCN components were used for styling and UI consistency.

🔐 Authentication & Wallet Integration We integrated Privy for secure wallet-based authentication. This enables users to log in using their crypto wallets, ensuring decentralized identity and privacy while interacting with the platform.

🧠 AI & Natural Language Processing To generate personalized quests and intelligent matchmaking, we integrated:

LangChain: This helped us manage prompt templates, memory, and agent flows for AI-driven quest creation.

OpenAI (GPT-4): Powers the AI agent that generates tailored quests and networking suggestions based on each attendee's interests, professional goals, and the theme of the event.

🗃 Backend We used Prisma ORM to manage our database, allowing for easy and efficient interaction with user profiles, quests, XP records, and event data. This provides a clean and type-safe approach to handling backend logic.

🔗 Smart Contracts Smart contracts were written to handle:

Event Creation: Recording event metadata on-chain

XP System: Storing and updating XP as attendees complete quests or form meaningful connections

Connection Tracking: Keeping an immutable on-chain record of verified networking interactions

These contracts are deployed on Celo, Base Sepolia, and Rootstock Testnet to ensure compatibility and experimentation across multiple EVM chains.

🦸 Superbase (Supabase) We used Supabase as a real-time layer for syncing off-chain data like live quest progress, pending connection requests, and real-time leaderboard updates. It acted as a bridge between the smart contract events and the frontend dashboard.

🧪 Notable Hacks We used LangChain agents in a clever way by combining real-time user data with AI-generated prompts to produce quests that felt personal and relevant.

We implemented an optimistic UI pattern with Supabase to show XP gains and new connections in real-time, even before the blockchain transaction was fully confirmed, improving user experience significantly.

AI matchmaking suggestions were enriched using clustering based on shared interests and historical XP behavior, mimicking a “reputation-based” networking suggestion system.