AWS Builder Center

Aegis GateKeeper: Building a Smarter World Cup Gate Experience

I built an AI assisted stadium gate system that helps volunteers make faster, more consistent access decisions while considering real-time queue and kickoff conditions. This article covers the idea, how I built it, what broke during development, and how I deployed it with AWS Amplify.

This summer, I wanted to build something that felt practical rather than just another AI demo. That idea became Aegis GateKeeper, an AI-assisted gate access decision system designed around the FIFA World Cup 2026 experience.
The problem is simple: at a busy stadium entrance, volunteers may have to make quick decisions about bags, tickets, cameras, phones, queues, and other access issues. When thousands of fans are arriving at the same time, even small decisions can create delays or confusion.
Aegis GateKeeper is designed to turn that process into a simple operational decision: ALLOW, REVIEW, or DENY.
The volunteer can provide a description of an incident and, where applicable, an image. The system also considers operational information such as current queue time and time remaining until kickoff. The AI then returns a structured decision with an action for the volunteer, an explanation of the reasoning, and a Spanish translation that can be communicated directly to the fan.
The goal wasn't to replace human volunteers. It was to give them a faster and more consistent decision-support tool when the gates are busy.

How I built it

The application is built with React, TypeScript, and Vite, with the project maintained in GitHub.
The decision engine uses a structured prompt containing the gate rules. These include clear-bag requirements, medical and childcare bag exceptions, ticket requirements, prohibited screenshots, camera restrictions, and dynamic gate routing based on queue time and kickoff time.
One feature I particularly wanted was dynamic routing. For example, if the queue is longer than 15 minutes and kickoff is less than 30 minutes away, the system can recommend redirecting a fan to an adjacent gate rather than simply following the default gate assignment.
I also added caching and validation around the decision process so repeated requests can be handled more efficiently and malformed image data doesn't unnecessarily reach the model.
For the AWS portion of the project, I deployed the production frontend using AWS Amplify Hosting. The source is maintained in GitHub, and I also created a GitHub Actions workflow to build the application and generate the deployment package.

What broke along the way

Honestly, getting the application deployed was almost as interesting as building it.
My AWS account initially had customer verification restrictions, which prevented me from creating some environments. I also ran into a Bedrock quota issue while trying to work with Amazon Nova Lite. The quota request had to be escalated to the AWS team, so I had to find another way to get the project deployed before the showcase deadline.
Then I hit a few smaller problems.
My first GitHub Actions build failed because  package.json  had invalid JSON formatting. After fixing that, the build failed again because  package.json  and  package-lock.json  were out of sync after an unnecessary dependency change.
Once those were fixed, the build finally passed.
The first Amplify deployment then returned a 404 even though Amplify showed the deployment as successful. The issue turned out to be the ZIP structure: the downloaded GitHub artifact contained another ZIP inside it. Uploading the actual deployment ZIP with  index.html  at its root fixed the problem.
That was a good reminder that deployment isn't just about getting the code to compile. The packaging and hosting configuration matter just as much.

What's next

Aegis GateKeeper is now publicly deployed on AWS Amplify, and the source code is available on GitHub.
This project started as an idea for making stadium access smoother, but the biggest lesson for me was the building process itself: prototype quickly, test the assumptions, expect things to break, and keep finding a path forward.
Huge thanks to Lewis Sawe for the inspiration this summer.
I’m excited to share Aegis GateKeeper with the Builder community and see what other builders would add to it next.
Any opinions in this article are those of the individual author and may not reflect the opinions of AWS.
Enjoyed reading this content? Let the author know!

Your likes, comments, shares, and saves help creators reach more builders.

Loading recommendations

Loading article