6,000 is the number that matters. During Lovable’s SheBuilds 24-hour global build on March 8, 2026, a solo participant used Lovable’s prompt-queue tooling and the event’s Anthropic API credits to launch ScuelaMIA, a searchable directory covering more than 6,000 Florida schools. The app pairs a conventional search and per-school profiles with an AI chat that answers questions by consulting the underlying structured dataset rather than returning generic web pages. The result is a practical prototype that makes enrollment deadlines, school types and locations queryable in one place.

The quick read is simple. ScuelaMIA proves that with the right tools a single developer can turn scattered web content into a structured, searchable public resource in a single sprint. The project was built during SheBuilds, a free, open 24-hour hackathon organised by Lovable and powered by Anthropic, and it shows what small teams can accomplish when tooling lowers the friction on data work.

How it was built in 24 hours

SheBuilds ran on March 8, 2026, with community events in more than 40 countries and a live stream hosted by Elena Verna from Stockholm. Participants were given a small bundle of resources meant to accelerate prototyping: each received $100 in Anthropic API credits and $250 in Stripe fee credits. One participant logged in from Miami and worked solo through the event, using the provided stack to prototype and assemble ScuelaMIA in a single session.

The build began with a focused scrape. The author started with Miami-Dade data and expanded to cover the entire state of Florida. Using the event's LLM-based tooling, the builder scraped and structured over 6,000 school entries. Each record was tagged with location, school type, profile content and enrollment deadline data. That canonical dataset was then used to populate both a search index and the app’s conversational agent, so questions like when a program accepts applications or whether a school offers language immersion can be answered by the chat bot drawing on the underlying records.

Lovable’s tooling played two practical roles. The LLM tooling handled the heavy lifting of scraping and initial structure, and Lovable’s prompt queue allowed the builder to run batches of prompts without manual intervention.

The prompt queue let thousands of prompts process in the background while the builder moved on to other tasks, which is how a single person could scrape, tag and publish so many records during a 24-hour window.

Mastheads

The build emphasised immediate utility over polish. ScuelaMIA shipped with three core features: keyword search across the dataset, individual school pages that expose the structured fields, and an AI assistant that queries the dataset to answer natural language questions. The conversational layer doesn't simply return links or web results. Instead, it consults the app’s records, so answers point to the same canonical facts parents see on a school’s profile.

What parents can do now

Parents and caregivers gain a different workflow when using ScuelaMIA. Instead of piecing together directories, PDFs and isolated reviews, they can search a single structured dataset. Each school profile lists geolocation, categorical tags such as charter, magnet, private, religious, STEM, arts and language immersion, plus enrollment deadlines when available. That makes direct comparisons easier and cuts the time spent chasing fragmented information.

The app also addressed a common technical hurdle. Lovable compiles projects as React single-page applications that render little to no HTML server-side. For a typical crawler, visiting such a site looks like a blank page. To ensure discoverability, the builder implemented a workaround that treats each school record as an individually discoverable item, enabling search engine crawlers to access content that a single-page app would otherwise hide.

That SEO-conscious choice matters for a directory aimed at families. If pages can't be crawled, users arriving via search will never see the profiles or the conversational answers. Making each school a separately indexable page both preserves the single-page app experience for users and opens the content to organic search traffic.

The personal motivation behind the build is also plain. The author says a partner repeatedly complained that "finding a school in Miami is impossible," and that annoyance anchored the scope and pace of the project. With the hackathon credits and the provided tools, the builder turned that complaint into a working site in a single development session.

ScuelaMIA isn't presented as an enterprise product. The project narrative and the feature set focus on practical, immediate value: searchable records, clear profiles and an AI assistant that answers fact questions by consulting the same records. The sources don't describe paid tiers, a rollout timetable or plans for expansion beyond the state-level coverage already captured.

There are two practical takeaways for anyone thinking about similar work. First, modern agent and scraping tools can compress what used to be a weeks-long ETL job into hours.

Second, single-page front ends still require deliberate steps to be indexable. The combination of LLM-based tooling for structure and a prompt queue for scale offered a working pattern during the hackathon that could be re-used in a production setting, with proper engineering hardening.

Related Articles

Built by a solo participant in Miami using hackathon credits, ScuelaMIA now lists searchable profiles for more than 6,000 Florida schools.

This article was created with AI assistance.