💡 Inspiration
As university hackathon organisers, we know the "organisational nightmares" all too well. We've lived through the chaos of last-minute changes, scattered spreadsheets, and the endless coordination required to get an event off the ground.
🚀 What it does
Hacksmith is an all-in-one management platform for hackathon organisers. It provides a central dashboard for all your event's crucial facts and figures, schedules, sponsors, mentors, participants, and more, allowing you to easily make changes and amendments when plans inevitably change.
The star feature is our integrated Agentic AI assistant, powered by Google Gemini. We train the assistant on your specific event data, turning it into a "hackathon rubber ducky." Organisers can ask complex questions about any aspect of their event and get instant, tailored answers pulled directly from their data.
Example: "What percentage of my overall budget have I spent so far?" or "How many attendees have dietary requirements and are my catering options fulfilling their needs?"
🛠️ How we built it
We chose a modern, scalable tech stack to build Hacksmith:
- Frontend: Next.js and React for a fast, component-based, and easily adaptable user interface.
- Database: Supabase to manage all of our event data in a robust, relational system.
- AI: Google Gemini API. We wanted to push beyond simple API calls. Our core technical challenge was deeply integrating Gemini with our Supabase backend, enabling the LLM to intelligently query large amounts of varied data from our database.
😥 Challenges we ran into
- Database Design: Our first major challenge was designing a flexible database schema. Mapping the complex relationships between primary and foreign keys (for sponsors, events, schedules, users, etc.) correctly was critical and took careful planning.
- Row Level Security (RLS): Our biggest blocker was with Supabase's RLS. We spent a significant amount of time debugging policies that were preventing our LLM backend from querying the database. Getting the permissions right so the AI could access the correct data without compromising security was a major hurdle.
🏆 Accomplishments that we're proud of
We are incredibly proud of building an AI-powered tool that solves a problem we are all personally passionate about.
While some of us had touched the Gemini API before, we had never used it on this scale. Fully integrating an LLM with a live database and a full backend, and achieving this in such a short time, is an accomplishment our whole team is proud of.
📚 What we learned
This project was a massive learning experience. Many of our team members learned Next.js and Supabase from scratch during this hackathon.
More importantly, we all learned how to integrate AI into a project in a meaningful way. Instead of a gimmick, our AI agent assistant is a core utility that provides genuine value to the user, making their planning process easier.
🔮 What's next for Hacksmith
We have a clear vision for the future of Hacksmith:
- Team Accounts & Collaboration: Currently, the platform is single-user per organizing group. Our top priority is to add team accounts so all organizers can work on the platform collaboratively, boosting the efficiency of the planning stage.
- Agentic AI Improvement: We want to improve our Agentic AI functionality to make our assistant even more powerful. The goal is to create an agent that can not only answer questions but also take action on the user's behalf (e.g., "Can you move the 'Intro to Git' workshop to 4 PM and notify all registered participants?"), truly streamlining the entire user experience.

Log in or sign up for Devpost to join the conversation.