From Prompt to Published: Ten Students, One Day and a Live URL

On May 23, 2026, ten students in grades 6 through 12 walked into a Denver classroom having never written a line of code. Six hours later, each of them had a working project, a live URL and a plan for a summer business. WWT sponsored lunch. The students did the rest.

Most youth tech programming ends the same way: a fun day, a certificate and nothing that survives the drive home.

The Vibe & Vision Venture Lab was built around a different finish line. Not did the kids have a good time; did they leave with something at a URL they could send to their grandmother.

Ten students, grades 6 through 12, one room in Denver, 10 a.m. to 4 p.m.

Dream it. Build it. Launch it.

The whole day runs on three words, and the order is the point.

Dream It. Before any tool opens, students figure out what they actually want to make. Not what's easy to build; what they want to exist.

Build It. The tool comes second, and it comes fast. Describe the idea in plain English, watch it become real, then start fixing what's wrong with it.

Launch It. The part almost every youth program skips. A project that lives on a laptop isn't launched. A project at a URL is.

Most workshops for this age group stop after Build. Stopping there is what makes it feel like a class instead of a launch.

The day, start to finish

The sequence was deliberate; nothing about AI tools until the students understood what they were about to use.

Welcome and house rules. Ground rules first, in a room with a twelve-year-old and a seventeen-year-old in it.

What is AI. Plain language. No jargon, no hype.

Using AI responsibly. Before touching the tool, not after. What it gets wrong, what it makes up and what remains the student's job no matter what the machine produces.

What is vibe coding. The concept that makes the rest of the day possible: describe what you want, and the code gets written for you.

Logging into imagi and Lovable. Accounts created, everyone on the same screen. imagi is what made this step possible; it's the layer that gave students safe access to Lovable, and free accounts to build on.

Lunch.

Writing their first prompt. The moment the abstraction becomes real.

Building their first game. Working through the [PIRATES?] curriculum in imagi, not a tutorial to copy, but a structure to build their own game inside.

Improving and debugging. The most valuable hour of the day, and the one nobody expects. The first version is always broken. Fixing it is where the actual learning happens, because a student has to figure out the gap between what they asked for and what they got.

Publishing their project. Live. On the internet. Real URL.

Remixing. Take what exists and push it further.

Showcase to parents. Every student presents what they made, to the people who dropped them off that morning.

Summer launch plan. What happens to this project after today.

Wrap-up.

What they actually built

The projects were not exercises. Students picked real businesses and built the thing.

A tenth grader built Tennis Swap, a youth marketplace for reselling gently worn gym shoes, so kids can earn toward their next pair instead of asking a parent for them. It has Google sign-in, magic-link email and guest browsing, so someone can look at listings without handing over an account.

The detail worth pausing on: it has report and block.

Nobody assigned that. A student building a marketplace where other teenagers message each other thought through what happens when someone is a problem, and built the controls before launch. Sign-in is only required to post, message, report or block; browsing stays open. That's a considered trust model, and plenty of shipped adult products don't get there.

Another student built Harps Garage, an appointment booking app for a car care business, where customers schedule maintenance and keep their vehicle information on file. It has full account creation, a logo and a business identity, not a placeholder name.

Two students, two entirely different business models. One built a marketplace where strangers transact with each other. The other built a service business with a booking flow and customer records. Neither is a tutorial project, and neither one looks like the other, which is what happens when the assignment is make the thing you want to exist instead of follow along.

Why debugging was the turning point

The temptation with a group this age is to keep everything smooth; pick a tool that always works, celebrate the first output, send everyone home happy.

That skips the lesson.

When a student's first build doesn't do what they pictured, they have to look at their own prompt and ask what was missing from it. That's the moment vibe coding stops being magic and becomes a skill. The AI did exactly what it was told. The gap was in the telling.

Every student in the room ran into that gap and worked through it. A twelve-year-old who has debugged their own prompt understands something about AI that most adults using these tools have not yet figured out.

The showcase

Presenting to parents changed the stakes in a way that a demo to instructors never does.

These students were not showing a worksheet. They were pulling up a live site, on a real URL, that had not existed when their parents dropped them off that morning.

Six hours. That's the whole gap.

What lunch actually bought

WWT sponsored lunch for the day, and the students were grateful in the specific, direct way that middle and high schoolers are grateful about food.

It's worth being plain about the size of that contribution: it was lunch. But for a free, full-day program serving grades 6–12, food is not a nice-to-have. A six-hour day without a meal is a shorter day, or a day some families can't send their kid to. Sponsored lunch is the difference between a program that runs 10 to 4 and one that has to cut itself in half.

Small sponsorships close real gaps in community programming. This was one of them.

What they left with

Free Lovable accounts, through imagi, the same platform used for professional prototyping, in their hands after the workshop ends, reached safely and at no cost.

A live URL hosting their project.

A certificate of completion, handed out at the showcase, in front of their parents.

The access problem nobody mentions

Worth naming, because it's the part that quietly determines whether a workshop like this can happen at all.

Professional AI build tools are not designed for middle schoolers. Getting a room of students into one safely, and without a credit card, is a real logistical problem, and it's the kind of problem that kills community programming before the curriculum is ever written.

imagi solved it. It gave students a safe path into Lovable and free accounts to build on, which meant the day could use the actual professional tool rather than a watered-down classroom substitute. Students left the room with working accounts on the same platform adults use.

That's the difference between a taste and a foothold.

A summer launch plan, what to do with it over the break, so the day is a starting line rather than a finish.

 

What scales

Ten students is a small number, and that's the argument, not the caveat.

The day required one facilitator, one room, laptops and a meal. No coding curriculum. No prerequisite. No technical background from anyone in the room, including the students. The entire structure, responsible use, then vibe coding, then build, then debug, then publish, then showcase, is portable to any city with a room and ten kids in it.

Grades 6 through 12 in the same session sounds like a problem and wasn't. When the tool is plain English, the differentiator isn't reading level or math background. It's how badly a kid wants the thing they imagined to exist.

Ten students. Six hours. Ten live URLs.

That's what happens when the finish line is launch instead of learn.