Miami's LIBA Is Mining Public Records With AI to Find Real Estate Deals Nobody Listed

MIAMI — The best real estate deal in South Florida might not be on any listing site. A small Miami startup called LIBA is betting that it is sitting in government records instead, and that artificial intelligence can find it first.

As Refresh Miami reported on Nov. 3, LIBA founder and CEO Akos Balogh recently zeroed in on an abandoned house in Fort Lauderdale that was not for sale. What flagged it was a detail buried in public records: unpaid property taxes. The house became an investment for LIBA Capital, the company’s real estate arm.

That deal is a small example of the bigger idea behind LIBA SI, the company’s AI platform for investors who want to spot opportunities before the rest of the market does.

Two Businesses Under One Roof

According to Refresh Miami, LIBA runs two related businesses. LIBA Capital invests the company’s own money in residential properties. LIBA SI licenses the underlying technology to professional investors.

The setup gives the startup a built-in test customer. LIBA Capital serves as the technology’s first user, trying out the platform’s recommendations on real acquisitions before outside investors rely on them.

Balogh came to the idea after his own frustrating hunt for multifamily properties. “We spent six months and I realized there has to be a system helping on that,” he told Refresh Miami.

He is not new to building companies. Balogh previously built Colle, a blockchain-powered luxury marketplace, and began rethinking what technology was for as enthusiasm for NFTs faded, the publication reported.

What the Platform Actually Tracks

Instead of relying on conventional listings, LIBA’s platform pulls from government records and other sources covering unpaid taxes, foreclosures, insurance, zoning and flood risks.

The numbers, as reported by Refresh Miami:

• Properties tracked: roughly 580 in Miami-Dade.
• Data points: more than 13,000, across 17 categories.
• Source mix: some 94% of the information comes directly from government records.

Co-founder Peter Liptak, a telecommunications infrastructure entrepreneur and multifamily investor, leads LIBA SI’s data network and infrastructure. “We built the data network the way you build a telecommunications network: every source tested, and every reading traced back to where it came from,” Liptak said.

That emphasis on traceability matters in a field where bad data can be expensive. Flood exposure, insurance costs and tax liens can change the math on a South Florida property dramatically, and investors need to know where each number came from.

Four AI Models, One Human Signature

Finding a deal is only half the job. According to the report, LIBA’s technology also helps investors decide when to walk away, weighing potential returns, financing costs, vacancy rates and other variables.

The startup runs four AI models, including OpenAI’s ChatGPT, Anthropic’s Claude, Google’s Gemini and xAI’s Grok, through a system it calls Consensus. The idea is to compare several models’ views rather than trust any single one.

People still make the final call. “All the decisions have to be principal approved,” Balogh said, adding that investors must document their reasoning so the platform can learn from past decisions.

A Focus on Everyday Housing

LIBA’s focus is not the glossy side of Miami real estate. Balogh framed the company’s work around affordable housing and the broader local economy. “It’s not the typical Miami story where you see the mega yachts and the fancy cars,” he said. “It’s more how normal people are living.”

That focus puts LIBA in a corner of the market that matters to a lot of local entrepreneurs: small and mid-sized multifamily buildings, where independent investors compete for deals without the research teams that big institutions have.

From Self-Funded to Venture Talks

So far, LIBA has funded development internally, according to Refresh Miami. The seven-person company is now onboarding outside multifamily investors through a subscription model and has opened discussions with venture capital firms.

Miami-Dade is the starting point. The company plans to expand across Florida and eventually nationwide, and it is already testing Las Vegas as another market. Balogh also sees the same decision-making technology potentially applying to other asset classes beyond property.

His goal is not modest. “We want to be a Miami-based startup, a Miami-based next AI unicorn,” he said.

Why It Matters for Miami Founders

LIBA is raising its profile at a busy time for South Florida startups. Refresh Miami’s analysis of PitchBook’s third-quarter data found that Miami-metro startups raised at least $1.14 billion across 117 deals in the third quarter and at least $3.14 billion through the first three quarters of 2026, putting the region on track for one of its strongest venture years.

Much of that money is going to AI companies. LIBA’s approach shows a different kind of AI startup: one that starts with a local, data-heavy problem, tests its product on its own deals, and only then sells it to others.

For founders, the lesson is practical. Public data that sits in county systems and city records is free to access, but turning it into decisions takes time and discipline. Companies that can make it trustworthy and traceable may find customers among Miami’s many small real estate investors, landlords and lenders.

The next test for LIBA is whether outside investors pay for the subscription, and whether venture firms back the push beyond Miami-Dade.