How I Built Brigitte, an AI Town Guide, in Replit Over a Weekend
Most people think you need a computer science degree to build a working app. I don't have one. I have Replit, a few AI tools, and a stubborn habit of building the thing instead of talking about it.
Last week I built an AI guide named Brigitte for our local BridgeFest here in Allegan, Michigan. She lives at brigitteallegan.com. She took about 36 hours of build time spread across a few days.
And the part I didn't see coming: the Chamber of Commerce and the City of Allegan want to keep her around.
Let me walk you through what I built, how I built it, and the bigger thing this points to for anyone who likes making tools.
Meet Brigitte the Bridge Guide
Allegan has a historic bridge. BridgeFest is the town event built around it. I wanted our booth to have something people would remember, so I made Brigitte a friendly animated guide who knows the town.
Here's what she can do right now:
- Answer questions about the festival, the town, and local businesses through an AI chatbot. The first 10 questions are free, then it's $9.99 for unlimited.
- Run a location-based scavenger hunt that walks people around the event.
- Tell dad jokes on demand. (I have a personal brand to protect.)
- Point visitors to a business directory so local shops get foot traffic.
She has an animated face that reacts while she talks. Small touch, big reaction from the people who saw it.
The Build: Replit, Clerk, and a Lot of Iteration
I'm a vibe coder. That means I build with AI doing the heavy lifting on the code while I steer the product and the experience. No traditional dev background. Replit is where all of it happens.
Here's the rough stack:
- Replit for the build environment and the hosting.
- Clerk for sign-in, so people could log in without me building login from scratch.
- Replit DB to store the data without standing up a whole separate database.
- A custom animated SVG for Brigitte's face.
The honest version of the timeline: 36 hours is not 36 smooth hours. There were dead ends. There were features that worked in my head and broke on the screen. I rebuilt the chatbot flow more than once.
But that's the real win with this kind of building. You get to test an idea with actual people in days, not months. I had a working product at a live event the same week I started.
What Happened at the Booth (the Honest Part)
I'll be straight with you. The booth traffic was lighter than I hoped. Fewer people stopped by the business booths than I expected, mine included.
That stung a little. You build something, you want a crowd around it.
Then the good news showed up. The people from the Chamber of Commerce and the City of Allegan saw Brigitte, and they got it right away. They want to use her to keep the community posted on events, Chamber happenings, and the big stuff like BridgeFest itself.
They also want to promote her out to local businesses and residents. That kind of feedback matters more than a busy booth. One real believer beats fifty browsers.
What's Next for Brigitte
The plan is growing past a one-weekend festival tool. A few directions I'm looking at:
- A live community events feed people can check year round.
- A spot for the Chamber to push updates out to local businesses.
- Smarter local search, so a visitor or a resident can ask "what's open right now" and get a real answer.
She started as a booth experience. She's turning into something the whole town can use. I'm working out what an official version looks like.
The Bigger Idea: A Tool Like This for Every Town
Here's the part I want builders to sit with.
Almost every town has a Chamber of Commerce. Most of them run their events on a patchwork of Facebook posts, email blasts, and a website nobody updates. A friendly AI guide that knows local businesses and local events solves a real headache for them.
That's a market. Not a someday market. A right-now one.
You could build it as a simple branded website for a small town. You could build it as a full app for a bigger one. Same core idea, scaled to the budget. A builder who packages this up could serve a dozen towns without reinventing the wheel each time.
I'm not saying it's easy money. I'm saying the need is sitting there in plain sight, and the tools to build it are cheaper and faster than they have ever been.
You Don't Need Permission to Build
The thing I keep coming back to: I built Brigitte because I wanted to, not because someone hired me to. The interest came after the build, not before.
If you've been waiting for the perfect plan before you make something, stop waiting. Make the rough version. Put it in front of real people. Let the feedback tell you what it wants to be.
Brigitte was a booth mascot on Friday. By the weekend she had a city interested in keeping her. Building first changes the whole conversation.
Want to see what a tool like this could look like for your town or your business? Visit brigitteallegan.com to meet Brigitte, or reach out and let's talk about what we could build together.
- Warren Schuitema, The AI Dad