A three-bedroom house in Miami sells for $4.2 million. All cash. The buyer on the deed: "Oceanview Capital Holdings LLC."

Who's behind it?

Whether you're a compliance analyst screening counterparties, a lender underwriting a CRE loan, an investor doing due diligence, or a journalist following the money — the question is the same. And in 2026, answering it requires a specific set of tools that most people don't know about.

At Zephira, we cover the data infrastructure behind corporate transparency — the registries, the APIs, and the gaps between them. This is the guide we wished existed when we started.

We'll walk through the exact process. Start with a property address. End with a natural person. And we'll be honest about where the trail goes cold.

What Changed in 2025

And why most guides on this topic are now wrong

March 2025 — FinCEN gutted the Corporate Transparency Act. An interim final rule exempted all U.S.-formed entities from beneficial ownership reporting. Only foreign-formed entities registered in the U.S. must file. A Delaware LLC that owns a $50 million tower? No filing required.

March 1, 2026 — FinCEN's Residential Real Estate Rule went live. Settlement agents must report all-cash residential transfers to entities or trusts. But it only covers residential, only non-financed deals, and reports aren't publicly searchable.

January 2026 — New York's LLC Transparency Act took effect, barely. Governor Hochul vetoed the expansion bill. It only covers foreign-formed LLCs registered in New York.

Find the Entity on the Deed

Every property has an owner of record at the county assessor. That's your starting point. Search by address on the county website, or use aggregated property APIs (ATTOM, CoreLogic, First American) for bulk lookups.

You're pulling four things: the entity name, the mailing address, the transfer date and price, and the grantor — who sold or transferred the property into this entity.

Look Up the Entity Where It Was Formed

Every LLC in the U.S. is registered with a state government. These registries are free and public. What they reveal depends entirely on which state.

Querying Registries with Code

Most states lack APIs. The OpenCorporates API aggregates 140+ registries. Free tier: basic entity data. Paid: officers and filings.

import requests

def search_entity(name, jurisdiction):
    # jurisdiction: 'us_fl' for Florida, 'us_de' for Delaware
    resp = requests.get(
        "https://api.opencorporates.com/v0.4/companies/search",
        params={"q": name, "jurisdiction_code": jurisdiction}
    )
    companies = resp.json()["results"]["companies"]
    if not companies:
        return None
    co = companies[0]["company"]
    return {
        "name": co["name"],
        "status": co.get("current_status"),
        "agent": co.get("agent_name"),
        "url": co.get("opencorporates_url")
    }

For UK entities, the Companies House API is free and far more powerful (Step 4). For broader international coverage, commercial providers like Bureau van Dijk, Global Database, and Moody's offer ownership data across hundreds of jurisdictions.

Walk the Ownership Chain

The registry shows the manager is another LLC — not a person. So you look up that LLC. And repeat. Until you reach a name or hit a wall.

Here's what a successful trace looks like:

Four layers. Three jurisdictions. Two countries. One person.

Now here's the more common outcome:

This is the most common result. The chain hits Delaware and stops. But there are side doors — that's Step 5.

First: what happens when the chain crosses international borders.

Foreign Registries That Give You More

If the chain leads to a foreign entity, you might be in luck. Several countries publish far more than any U.S. state.

UK Companies House — The Gold Standard

The UK requires every company to disclose Persons with Significant Control — anyone with 25%+ ownership. Free API. Full name, nationality, country of residence, nature of control.

def get_uk_owners(company_number, api_key):
    # Free key: developer.company-information.service.gov.uk
    resp = requests.get(
        f"https://api.company-information.service.gov.uk"
        f"/company/{company_number}"
        f"/persons-with-significant-control",
        auth=(api_key, "")
    )
    return [
        {"name": p["name"],
         "nationality": p.get("nationality"),
         "control": p.get("natures_of_control", [])}
        for p in resp.json().get("items", [])
    ]

Other Countries with UBO Data

France — via the INPI. Luxembourg, Netherlands, Austria — UBO registers for entities with "legitimate interest." Canada — federal beneficial ownership registry since 2024.

Opaque Jurisdictions

BVI, Cayman, Panama — ownership data held by authorities only. Public data won't get you there.

Workarounds That Actually Work

This is what most guides skip. Here are five side doors when the state registry gives you nothing.

The Delaware Dead End — Five Ways Around It

Nominee Directors

You traced the chain to a name — but it appears on 300+ unrelated entities. That's a professional nominee. Search their name across registries. If they show up on dozens of unrelated companies across multiple countries, they're paid to be the legal front. The actual controller's name is in the operating agreement — a private document.

Series LLCs

Deed says "XYZ Holdings LLC - Series 14" but only the parent is registered. Strip the series designation. Look up the parent. Its officers control all sub-series.

Land Trusts

Not business entities. Don't appear in registries. Trustee may be on the deed. Property tax records sometimes list a contact. But trust agreements are private — you usually need cooperation or legal process.

The Honest Summary

What's actually achievable — scenario by scenario

The Decision Tree

The full process, as a series of questions

Doing This at Scale

The manual process works for one investigation. For a portfolio, a lending pipeline, or a CRE analytics platform — you need three data layers:

Property data at scale — aggregated county records via ATTOM, CoreLogic, or First American. Multi-jurisdiction registry data — officer and shareholder information across U.S. and international registries. Ownership graph traversal — walking chains programmatically until you reach a natural person or a documented dead end.

No single provider covers everything. The right approach combines public data, commercial data, and investigative techniques — and stays honest about where each one runs out.