TL;DR
- Use a business registry to establish legal identity, jurisdiction and filing status. It is not a substitute for a commercial company dataset or proof that an account is a good prospect.
- Companies House is the first source to evaluate for UK registered entities. OpenCorporates is useful when a workflow needs company records across jurisdictions through one interface. GLEIF adds an open international identifier and reference data for entities with an LEI.
- Match on registration number and jurisdiction when available. A company name alone should produce candidates, not an automatic CRM overwrite.
- Include searches, detail calls, ambiguous matches and review labour in the budget. The number of input accounts is rarely the number of API calls.
What a registry record proves
The name in a CRM is often a brand, group or domain owner. The entity signing the agreement may be a different company. A business registry API helps connect those representations to a legal record: registered name, registration number, jurisdiction, status and available filings.
That is valuable for account matching and procurement. It is less useful for finding a current head of operations or estimating whether a business needs your offer. Keep the registry record alongside commercial evidence, rather than replacing the entire account with whatever the registry returns. Our company data provider guide covers that broader research purchase.
An active registration is not a guarantee of trading activity, solvency or data accuracy. A registered address may be an accountant's office. A filing date is not necessarily the date the underlying business event happened. These distinctions matter when an agent tries to turn a legal record into an account narrative.
Compare providers on the jurisdictions you need, source provenance, identifiers, update behaviour and permitted downstream use. “Global” is too vague for a team whose difficult accounts are regional subsidiaries in three specific markets.
Companies House for UK legal entities
The Companies House Developer Hub provides access to public company information, including company details, officers and filing-related interfaces. It is a strong starting point for a UK verification workflow because the authority is the register itself. Use the search endpoint to obtain candidates, then a company number to retrieve the selected record.
Companies House documents a limit of 600 requests in five minutes. Treat that as an API constraint, not a throughput target that every worker can use independently. Multiple processes using the same credentials need shared quota control. A failed or throttled request must remain pending, rather than marking the company absent.
The public API is a practical low-cost source for UK registered entities, but the API is not a ready-made account matcher. Your system must still reconcile the trading name and domain with the legal company. Nor does a UK register solve verification for companies registered elsewhere. Use local authority sources for those markets or a provider that covers them.
Choose this route when UK legal identity is the problem and you can operate the matching logic. It is less suitable as the sole source for a multi-country business intelligence product. Direct data access removes a commercial aggregator from the chain; it also leaves your team responsible for jurisdiction-specific fields and interpretation.
OpenCorporates for cross-jurisdiction research
OpenCorporates offers API and bulk access to its cross-jurisdiction legal-entity dataset. Its documented purchase choices separate a self-service API from bespoke bulk delivery. That distinction matters: a thousand selective checks and a recurring copy of several country datasets are different jobs.
Prices observed on 6 October 2026 are in pounds sterling. Essentials lists £225 monthly, up to 500 API calls a month and 200 a day. Starter lists £660 monthly, up to 2,500 calls a month and 500 a day. Annual totals are £2,250 and £6,600 respectively. Basic lists £1,200 monthly or £12,000 annually, with 5,000 monthly calls. These are access packages, not a guarantee of that many matched accounts.
The value is a common interface and documented source trail across jurisdictions. The limiting factor is still coverage of the exact register, fields and observation dates needed for your accounts. A common schema does not make every country's filings equally current or equally detailed.
Read the applicable self-service API terms for the intended use. Public or share-alike access and a commercial licence are different arrangements. The licence also describes restrictions on data flagged as restricted for marketing. The current commercial pricing page advertises internal and external use; the specific agreement remains the authority for an agency or embedded product. Do not infer unrestricted resale or unrestricted marketing from a plan label.
Evaluate OpenCorporates when maintaining separate registry integrations is more expensive than the subscription and its sources cover your target markets. For a UK-only internal lookup, the direct register is a simpler starting point. For high-volume research, size bulk access before trying to force the workload through a small API allowance.
GLEIF when an LEI resolves the identity
GLEIF's API supports access to Legal Entity Identifier reference data. An LEI is useful when both systems know the same identifier: it removes much of the ambiguity in comparing names and addresses across jurisdictions.
GLEIF also publishes concatenated files daily. That creates a different deployment option from making a remote lookup for every account. A local indexed copy can support repeated reconciliation, provided the team tracks the snapshot and subsequent refreshes.
LEI coverage is not the universe of B2B companies. A company without an LEI does not become invalid or commercially uninteresting. Distinguish the entity's legal status from the LEI registration status: a lapsed identifier is not automatically a dissolved business. Ownership relationships also need their reporting context; absence of a parent relationship is not proof of independence.
Use GLEIF as a high-value identity layer for relevant financial, multinational and other LEI-holding accounts. Do not choose it as the only discovery database for a local-business campaign. Its open data reduces licence acquisition cost, but parsing, storage, reconciliation and exception handling remain operating costs.
Build an acceptance rule before matching
Suppose “Northstar Logistics” arrives with a UK domain and no company number. The search returns an active company, a dissolved company and a similarly named Scottish subsidiary. A first-result policy might attach the right brand to the wrong contracting entity.
Use a staged rule:
- If a registration number and jurisdiction are supplied, check that record directly and verify that the accompanying name is plausible.
- If only a name is supplied, generate candidates. Use jurisdiction, address and a source connecting the website to the legal entity to narrow them.
- If two entities remain plausible, retain both candidates and route the match for review. Do not overwrite the account's company number.
- Record the selected legal entity separately from the commercial group and any operating subsidiary.
The output should include the provider, source record, retrieval time, registration number, jurisdiction and reason for accepting the match. Keep conflicting trading names visible. A record can be correctly identified but fail your commercial policy because it is dormant, outside the territory or the wrong group member.
Use the ABM Account Scoring Template to keep commercial qualification downstream of identity verification. The legal match establishes who the account is; your scoring policy decides whether it belongs in the programme.
Price a five-jurisdiction verification job
Take an illustrative job of 5,000 company records across five jurisdictions, with 1,000 in each. Assume 800 per jurisdiction resolve and require a detail request, while 200 remain ambiguous. That produces 9,000 API calls across the programme: 5,000 searches plus 4,000 details, before pagination, retries or filings. The resolved/ambiguous split is a scenario, not a measured provider coverage rate.
The observed OpenCorporates Basic allowance of 5,000 monthly calls cannot cover all 9,000 in one month. A team could spread the job over two months under the appropriate daily limits, obtain a larger API package or request bulk access. Two monthly Basic payments at the observed GBP1,200 rate total GBP2,400; this assumes the current monthly offer and permitted use remain available, rather than a discounted annual commitment or negotiated quote.
Reviewing the assumed 1,000 ambiguous cases at three minutes each takes 50 hours. At an illustrative loaded GBP40 an hour, review costs GBP2,000. Access plus review is therefore GBP4,400, or GBP1.10 per resolved entity across the assumed 4,000 accepted matches. Integration, tax, documents and additional licence requirements are excluded.
Do not assume every jurisdiction produces the same fields or matching yield. Inspect each country separately: available identifiers, register source, status definitions, update dates and licence restrictions can change the acceptance rule. A cheap common interface is useful only when it covers the legal entities and evidence your five-country programme needs.
If the job is UK-only, a direct Companies House workflow changes the access component, but not the need to match and review correctly. If relevant inputs already carry valid LEIs, a GLEIF-based match may remove many name-search steps. The cheapest source for one jurisdiction need not be cheapest for a team maintaining five different integrations.
For a recurring programme, compare annual commitment with workload concentration. A plan with sufficient monthly calls can still block a one-day import through its daily cap. Spread the job deliberately or buy the necessary capacity; don't silently defer half the verification and present the import as complete.
Put the result into the account system
Write registry fields into dedicated fields with provenance. Preserve the seller's trading name and account hierarchy unless the accepted update policy explicitly changes them. Legal identity, corporate group and lead routing are related decisions with different owners.
For integration entitlements and data delivery, use our B2B data API guide. For the next step in Salesforce, the lead-to-account matching guide explains how identity should feed the routing decision.
Our buying preference is direct official data for a well-bounded jurisdiction, OpenCorporates when the common interface and covered markets justify its charge, and GLEIF as an additional identifier layer where LEIs are relevant. None should be sold internally as proof of buying intent. The useful outcome is fewer wrong entities entering research, routing and commercial decisions.
FAQ
Is a company domain enough to identify the legal entity?
Usually it is a useful clue, not a unique legal identifier. A group can operate several entities under one domain, and a brand can differ from its registered company name. Retain the domain-to-entity evidence and use registration number plus jurisdiction for subsequent checks.
Does an active registry status mean the company is trading?
It establishes the recorded status in that register. It does not necessarily establish current trading activity, financial health or fit for your offer. Combine it with appropriate current commercial evidence rather than adding automatic buying-intent points.
Can an agency share registry results with its clients?
That depends on the dataset and applicable licence. Internal access, external distribution and marketing use are separate questions. Scope the intended client deliverable in the agreement; don't assume that a paid API key grants unrestricted redistribution of the underlying dataset.
Should an ambiguous match be rejected?
It should be withheld from automatic acceptance. Keep the candidates and explain what evidence would resolve them. A short manual review can be cheaper than merging the wrong legal entity into an existing account and repairing routing, reporting and contract records afterwards.
