Corporate Registries API: Official Company Data From Government Registries

A company's legal identity — whether it's actually registered, who's legally accountable for it, whether it's dissolved or in good standing — isn't a matter of opinion or inference. It's a matter of public record, published by a government register in whichever country the company is incorporated in. The problem is that "public record" means a different interface, in a different language, with a different data shape, for every country you need to check. A corporate registries API turns that patchwork of official sources into one consistent, structured call per country, without requiring you to learn eleven government websites to run one due-diligence check.

What corporate registry data actually covers

Before comparing ways to get it, it's worth being specific about what a registry actually publishes, since it varies meaningfully by country: registration status and incorporation date at a minimum, and often officers and directors, beneficial ownership, filed financial statements, authorisations, or adverse entries like bankruptcy proceedings — whatever that specific jurisdiction's register happens to hold.

vs. a commercial company-intelligence database

A commercial database aggregates company information from many sources into one composite profile, which is convenient when you want breadth from a single call. The tradeoff is that a composite profile is a step removed from the primary source — you're trusting the aggregator's freshness and interpretation rather than reading the register directly. For a question where the legal fact itself is what matters — is this entity active, who is legally accountable for it — reading the primary government source directly is the more defensible answer, especially for compliance work where "where did this fact come from" is itself part of the requirement.

vs. searching each government register by hand

The free option is opening each country's official register and reading the entity page directly. It works for checking one company before a single deal, and falls apart the moment a due-diligence process needs to run across a portfolio, or the same check needs to repeat on a schedule — a dozen countries times a dozen counterparties is not a process anyone sustains by hand for long, and every register is in a different language with a different page layout to learn.

vs. building a scraper per register

The engineering option is writing a scraper against each government register's own pages. That means taking on ongoing maintenance for every register you add — pages change, and a government website has no obligation to keep a stable structure for something scraping it. It also means solving the same problem, government registry structure, over and over, once per country, instead of once.

A structured API reading directly from official registers is the fourth option: one consistent request shape per country, returned as typed JSON, with the primary source URL attached to every result.

Corporate registries in NeuralVerge's catalog

Each country's register is its own source, built around what that specific register actually publishes, but reachable through the same request shape and response envelope as everything else in the source catalog:

• —Companies House (United Kingdom) — registration status, incorporation date, registered office, SIC codes, current officers, and persons with significant control in one call, plus separate sources for the full officer history and the complete filing history.
• —Ariregister (Estonia) — one of the more open registers in Europe, publishing shareholders and ultimate beneficial owners alongside the registration basics, which most countries treat as private.
• —CVR (Denmark) — the rare register that publishes filed financial statements: gross profit, profit before tax, result, equity, and total assets per year, alongside ownership and headcount.
• —YTJ (Finland) — the joint register run by Finland's Patent and Registration Office and Tax Administration, including which official registers — trade, VAT, prepayment — an entity is actually entered in.
• —INPI (France) — France's consolidated national register, naming the people with legal authority to bind the company alongside its establishments and filed documents.
• —KBO/BCE (Belgium) — the Crossroads Bank for Enterprises, notable for making group relationships between linked entities explicit rather than something you infer.
• —KRS (Poland) — the National Court Register, which carries adverse entries most registers don't: prohibitions on individuals, bankruptcy proceedings, and permits, alongside the standard registration data.
• —Czech Business Register search — a name- or identifier-based search returning structured candidates from a register with no English interface of its own.
• —LEI Lookup (global) — the one identifier that means the same entity in every jurisdiction, with a separate search source for resolving a company name to its code first.

Every one of these returns through the same outer response shape, so a due-diligence pipeline built against one register's output doesn't need a rewrite to add the next country.

A worked example: a cross-border KYB check

Take a concrete, illustrative case: onboarding Acme Oy (Finland) as a vendor, where the compliance requirement is confirming the entity is legitimate, active, and legally accountable to a named person before a contract is signed.

1. —A YTJ lookup against Acme Oy's Finnish Business ID confirms it's registered, active, and entered in the trade, VAT, and prepayment registers — the practical answer to whether it can be traded with normally.
2. —Because Finland's registration system holds registration facts rather than financial statements, a CVR-style check isn't applicable here — but if the counterparty had been Danish instead, the same due-diligence step would also surface filed financial statements in the same call.
3. —A LEI Lookup search against the company name checks whether Acme Oy holds a global Legal Entity Identifier — common for companies operating across borders — which, if found, becomes the identifier used to reference the entity in any cross-border reporting the deal requires.
4. —The compliance record for the vendor file is now built from three primary-source facts — active registration, correct register entries, and a global identifier where one exists — each with the government source URL attached, rather than a single composite score from a commercial database with no visible primary source behind it.

A worked example: mapping beneficial ownership across two jurisdictions

A second, distinct case: a due-diligence review needs to establish who actually owns and controls a UK entity that a corporate structure chart shows as linked to an Estonian parent.

1. —A Companies House company lookup against the UK entity's company number returns its persons with significant control directly — the people or entities that own or control it, with the nature of that control.
2. —If the parent shown in the ownership chain turns out to be the Estonian entity, an Ariregister company lookup against its Estonian registry code returns its own shareholder list and ultimate beneficial owners — one of the few European registers that publishes this directly rather than requiring a paid extract.
3. —Comparing the two results traces the ownership chain from the UK operating entity up to the natural persons who ultimately control it, using two primary sources rather than a single composite ownership graph whose underlying sourcing may not be visible.
4. —Because each lookup only bills on a resolved entity, tracing a chain that turns out to be a dead end at one register — an entity that doesn't exist there, or a register that doesn't publish shareholders — costs nothing for that step.

A worked example: monitoring a UK supplier's filing history for risk signals

A third, distinct case: a credit review team wants to catch a supplier's financial distress early rather than learning about it from a missed invoice.

1. —A Companies House company lookup against the supplier's company number, run on a recurring schedule, tracks its current status — active, dissolved, or in some stage of insolvency — as a first-pass signal.
2. —A Companies House filings lookup against the same company number returns the filing history with dates and links; the team compares each run against the last to catch new filings — a charge, a change of registered office, or accounts filed later than the prior year's pattern — as they land on the public register.
3. —Because filings are the register's own paper trail rather than a third party's interpretation of it, a pattern like "accounts consistently filed at the last permitted moment" is visible directly, months before it would show up in a commercial credit score built on a slower refresh cycle.
4. —Because each scheduled check only bills on a resolved company, running this across a full supplier list on a recurring schedule is a predictable, bounded cost rather than an open-ended monitoring subscription sized for a much broader use case.

Integration modes: pipeline call vs. agent tool call

A registry lookup is an ordinary API call, reachable directly from backend code or exposed as a tool an agent can call mid-task — the same choice that applies across the rest of NeuralVerge's source catalog. Calling it directly fits a compliance pipeline that already knows a registry check belongs at a specific point in onboarding, or on a fixed monitoring schedule, as in the worked examples above — the call happens because the workflow says it happens, at a point a developer chose in advance.

Exposing the same lookups as tools fits a due-diligence agent that's reasoning through a compound question — verify this counterparty, trace its ownership, check for adverse entries — and needs to decide for itself, mid-task, which specific register to check next based on what it's already found. An agent working a cross-border ownership chain, for instance, might not know in advance that a UK entity's parent turns out to be Estonian until the first lookup returns that fact — at which point it reaches for the Ariregister tool because its own reasoning surfaced the need, not because a developer scripted that specific sequence. NeuralVerge's registry sources are reachable both ways, over REST or MCP, returning the same response shape regardless of which interface triggered the call.

Where the catalog's coverage has real limits

Registry coverage is genuinely uneven across countries, and it's worth being direct about that rather than implying otherwise. Some registers, like Estonia's and Denmark's, publish unusually rich detail — shareholders, beneficial owners, filed financials — because their national policy favors openness. Others, like Poland's, publish adverse legal entries most registers omit. And several major economies aren't yet covered at all, which means a global KYB process built entirely on this catalog will hit real gaps outside its current country list — worth checking the full data source catalog against the specific countries a given compliance process actually needs before assuming full coverage.

Where teams use corporate registry data

• —Know-your-business and vendor onboarding. Confirming a counterparty is registered, active, and legally accountable to a named person before signing anything.
• —Beneficial ownership screening. Tracing who ultimately owns and controls an entity, using registers that publish shareholders and controlling parties directly.
• —Credit and risk review. Watching for adverse entries — bankruptcy proceedings, prohibitions, late filings — that a commercial database may not surface as directly as the primary register does.
• —Cross-border entity resolution. Using an LEI to keep the same legal entity from becoming several disconnected rows across a global account or vendor list.
• —Broader due-diligence workflows. Registry facts pair naturally with AI research for the parts of a diligence question a registry alone can't answer — recent news, funding activity, reputation — combined into one investigation rather than run as separate, disconnected checks.

What to check before you commit to a corporate registries API

• —Does it read the primary government source, or a commercial aggregation of it? For a compliance question, being able to point to the actual register the fact came from is often part of the requirement, not just a nice-to-have.
• —Does coverage actually include the countries your process needs? Registry coverage varies a lot by provider — check the specific country list against your actual requirement before assuming broad coverage.
• —Does it bill on a miss, or only a resolved entity? Running checks against a portfolio of counterparties of unknown registration status is a different cost proposition depending on the answer.
• —Can you resolve a name to a registration number, or does every register require the number up front? Most records don't arrive with a government identifier already attached — check whether name resolution is built in or a separate step.
• —Does the response fit into the same pipeline as your other research and enrichment calls? A registry lookup that returns a different shape from everything else in a compliance pipeline is one more parser to maintain.

Running a real counterparty — ideally one you already know the correct answer for — through a candidate source is a faster way to judge accuracy and coverage than reading a features page.

Frequently asked questions

Is this data pulled from an official government source?

Yes. Each source reads a specific country's public government or court-maintained register — Companies House in the UK, the e-Business Register in Estonia, and so on — rather than a commercial database compiled from other sources.

What if a company doesn't have a matching entity on a given register?

The lookup returns nothing to report rather than a guess, and isn't billed. A company not resolving on one country's register is expected if it's not actually registered there — that's a real answer, not a failure.

Can I search by company name instead of a registration number?

It depends on the register. Some, like the Czech Business Register, expose a name-based search directly. Most others take the registration number itself, so resolving a name to a number first — via AI research or a name-based registry search — is a separate step for those.

Does every country's registry return the same fields?

No, and that's by design rather than a limitation. Each country's register publishes different information — Denmark's CVR includes filed financial statements, Poland's KRS includes bankruptcy and prohibition entries, and Estonia's Ariregister includes shareholders. The response schema reflects what that specific register actually holds.

How current is registry data?

It's read at the moment you make the request, so it's as current as the register itself. Most registries update when a company files something, which for most fields means a handful of times a year rather than continuously.

Can I run these checks on a recurring schedule to catch changes?

Yes — a registry lookup is an ordinary API call, so running it against the same entity on a schedule and comparing each result to the last is a normal way to monitor for changes like a new filing, a status change, or a new officer, rather than a one-time onboarding check.