Open source company data for commercial use: registries and licences

Open sources of company data can carry a commercial product, each on the terms its publisher states, and those terms differ more than the word open suggests. This post reads the publisher's own page for five sources, the SEC, Companies House, GLEIF, Wikidata and OpenCorporates, and reports what each says on commercial use, attribution and share-alike. After it you can sort the sources behind a product by their conditions and take a short list of questions to counsel. Every text was read on 7 October 2026. None of this is legal advice.

Which open sources of company data allow commercial use?

Commercial use of open company data rests on the terms each publisher states. CC0 1.0, applied by GLEIF and Wikidata, waives the publisher's rights for any purpose, commercial included. The Open Database License (ODbL), named by OpenCorporates, includes commercial use with notice and share-alike conditions. The SEC and Companies House publish statements on reuse; we found no data licence named.

Two of the five pages showed no date of their own, so record the day you read each text.

Four kinds of terms behind the word open

A statement on public information. The SEC's text permits copying and redistribution of what its site presents, asks users not to use the SEC seal or the site's logos, and requires its licence for the trade marks SEC and EDGAR in a related business's trade name, trade mark or domain name. Alongside it stands the US Copyright Act: section 105(a) says copyright protection under that Act is not available for any work of the United States Government, which section 101 defines as a work prepared by an officer or employee of the government as part of that person's official duties. Questions for counsel: does a document a company files fall within that definition, what does the SEC's statement add for it, and what applies outside the United States?

A public-domain dedication. CC0 1.0 Universal is a waiver with a licence behind it. Section 2 waives copyright and related rights, database rights among them, for any purpose, commercial included. Section 3 grants an unconditional, royalty-free licence where the waiver cannot take effect. Section 4 keeps three things outside: the affirmer's trade marks and patents, any warranty, and responsibility for clearing other persons' rights in the work. Whose other rights may sit in a record is left with you.

An attribution licence. The Open Government Licence v3.0, which GOV.UK content carries except where otherwise stated, lets you copy, publish, distribute, transmit and adapt the information and exploit it commercially, provided you acknowledge the source with the provider's attribution statement or the default one and, where possible, link to the licence. Among other things it leaves out personal data, logos and third-party rights the provider cannot license, and we found no share-alike condition in it. The Companies House guidance sits on such a page, yet its words on register data name no licence: ask counsel which text governs the files you download.

A share-alike database licence. ODbL 1.0 says in section 3.1 that its rights explicitly include commercial use and exclude no field of endeavour. Section 2.4 limits it to the database: rights in individual contents, other than database rights or contract rights, lie outside it. Section 4.4 allows a Derivative Database to be used publicly only under the ODbL, a later version or a compatible licence, and counts it as used publicly when a Produced Work made from it is. Section 4.6 has you offer recipients the whole Derivative Database, or a file of every alteration or of the method of making them, in machine-readable form. Section 4.5 draws two lines a build can use: internal use of a Derivative Database is not public use, and a Collective Database, the original unmodified within a collection of independent databases, need not be licensed under the ODbL, though the licence still applies to the original inside it.

Which free, open company registry APIs license their data for commercial use?

Three official registers publish company data free of charge with a stated position on reuse. GLEIF provides LEI data free under CC0 1.0. The SEC's EDGAR APIs need no key, and its site says information may be copied and further distributed. Companies House offers free data products and states that it imposes no rules on use.

Reuse and access are separate terms, and a build meets both.

The Companies House guidance notes that the register holds data about companies and their officials: you are responsible for complying with data protection, copyright and other legislation, and should check the legal requirements before you publish. Which fields of a download name a person is an early question for counsel.

OpenCorporates, Companies House bulk download and US SEC companyfacts: how each source licenses company data for commercial use

OpenCorporates licenses its database under the Open Database License, with a credit link and share-alike, and names non-share-alike API terms for commercial applications. Companies House states in its data products guidance that it imposes no rules on how register information is used. The SEC says information presented on sec.gov may be copied and further distributed without its permission.

OpenCorporates. Its terms describe a database sourced from government and other sources. Information that combines its database with yours is to be published under the same licence, unless you agree a non-share-alike licence with the company. The terms offer free website access for personal use only, and share-alike API keys to Permitted Users (journalists, NGOs and academics conducting public benefit research) with evidence of contributing data or services back to the community under an open licence. They state that corporations and financial institutions are not Permitted Users, and name non-share-alike API keys for commercial or proprietary applications. Section 2.4 of the ODbL suggests the question to ask: where a record began in a government register, which text governs that record?

Companies House bulk download and SEC companyfacts. The snapshot comes as ZIP files of CSV, compiled to the end of the previous month. One call to the company facts API returns all the company concepts data for one filer, and a nightly bulk file holds the company facts data. We found no licence on the download page or the API page, so the two statements above are the texts to keep on file.

Wikidata business entities, OpenStreetMap and the ODbL: how an open dataset licenses commercial use of company data

Wikidata releases all its structured data into the public domain under CC0 1.0, whose waiver names commercial purposes. The Open Database License also includes commercial use, in section 3.1, and attaches notice and share-alike conditions to public use. OpenStreetMap's own copyright page names the ODbL as the licence of its data and requires credit to OpenStreetMap and its contributors.

Wikidata. The dedication covers the main, Property, Lexeme and EntitySchema namespaces; text elsewhere on the site is under CC BY-SA 4.0. For a company, the project's list of company properties includes the Legal Entity Identifier, ISIN, Central Index Key, Companies House company ID and OpenCorporates organization ID, beside attributes such as headquarters location, inception and employees. Where those identifiers are filled, an item is a crosswalk between the registers above. The data access page lists a query service, APIs and database dumps. The licensing page adds that the project avoids hosting data that lacks a CC0 licence.

A map dataset under the ODbL. A company product meets one when it places companies on a map. A map image or a screen can be a Produced Work, which carries a notice under section 4.3, and a table of your companies matched to map features and passed to customers can be a Derivative Database. Ask counsel whether the matched table is a Derivative Database or part of a Collective Database.

What open data leaves you to build

Three jobs remain once the licence question is answered.

• Matching. Each source keys its records its own way: a CIK at the SEC, a company number at Companies House, an LEI at GLEIF, an item at Wikidata. Where no shared identifier is recorded, the join is a match on name and country: entity resolution, with an error rate to measure. The identifiers that make company data joinable sets out what each key names.
• Dating. A snapshot replaces the one before it. To say what a record held on a past day, keep each delivery with its date, the basis of point-in-time data.
• Upkeep. Rate limits, an unsupported snapshot and terms pages without dates are yours to watch. Keep a register of provenance: for every field, its source, the governing text, its version or date, and the day you read it.

Worked example: Acme Robotics builds a supplier directory on open sources

This example is illustrative. Acme Robotics is an invented company and every figure is invented.

Assumptions:

• Acme Robotics sells a supplier directory to its customers and tracks 2,000 suppliers in it.
• 400 suppliers are SEC filers with a CIK. 900 hold an LEI, and 300 of those 900 are among the 400 filers.
• 700 other suppliers are UK companies with a company number. The rest carry no identifier.
• The directory has 12 columns: 4 from SEC data, 2 from LEI data, 2 from Wikidata, 3 from a database licensed under ODbL 1.0, and 1 score of Acme's own.
• Counsel has yet to review the build.

The working:

1. Matching. 400 + 900 = 1,300, less the 300 counted twice, gives 1,000 suppliers that resolve on a CIK or an LEI. The 700 UK companies bring the total to 1,700. That leaves 2,000 less 1,700, or 300 suppliers, 15% of the directory, to match on name and country.
2. Columns by terms. 4 + 2 + 2 = 8 columns come under a public-information statement or CC0 1.0, 3 under the ODbL and 1 is Acme's: 8 + 3 + 1 = 12. So 3 of 12 columns, 25%, bring the conditions of section 4.
3. Outputs. Each output is sorted on its own.

Acme has three designs to put to counsel: export the 9 columns outside the ODbL (8 + 1) and leave the ODbL database unmodified beside them as part of a collection; take a non-share-alike licence from the publisher; or replace the 3 columns with data licensed under an agreement that grants redistribution.

A checklist before a product is built on an open source

1. Find the publisher's own terms page and record its title, its version or date, and the day you read it. A catalogue's licence label only points to that page.
2. Write three answers for each source in the text's own words: commercial use, attribution, share-alike. Write "none found" where that is the result.
3. Keep terms on access, such as rate limits, user agents and keys, apart from terms on reuse.
4. Sort every output: internal, shown on a screen, or passed on as rows. For a database licence, decide whether the build alters the database or holds it unmodified in a collection.
5. List what the text leaves outside: third-party rights, personal data, trade marks, any warranty.
6. Take the open points to counsel as questions, with section numbers, and read each text again when a design changes.

Where licensed data fits

We license Fokals data by written agreement, for internal use, embedding in a product or redistribution. What a data licence covers sets the three uses side by side. For the three jobs above:

• Matching. Every company has one stable Fokals company ID. A listed company carries its ticker, MIC, ISIN, LEI and share-class FIGI, and its SEC CIK where it has one.
• Dating. Every record carries the time it was observed, and daily and weekly datasets are written once, after the period closes, and never revised.
• Upkeep. The data is refreshed daily and delivered by REST API or as bulk files.

Attribution travels in the delivery. Web Traffic, in the announcements and scale product, is delivered with attribution: each row carries the attribution text for its monthly traffic tier, defined in the data dictionary. A bulk export can come as a package of files with a manifest naming its period, versions and licence. The data is collected from first-party company sources and public records and processed in-house, as the sourcing statement sets out.

Frequently asked questions

What this page says about the products it names was checked against their public documentation on 7 October 2026. Product and company names are trademarks of their owners. Fokals is not affiliated with them or endorsed by them.