Companies House
Companies House keeps the statutory record of every company registered in the UK, millions of them, live and dissolved. For each one you can see its registered address, incorporation date, company status, directors and officers, the people with significant control (the humans who ultimately own or control it), the full filing history, and the published annual accounts.
There are three ways in, and which you pick matters. The website is fine for one-off human lookups. The REST API (free key required) returns `JSON` and is the right tool for pulling structured details on specific companies on demand, including a document API for the filings themselves. The bulk data product is a monthly `CSV` snapshot of basic company data for the whole register, ideal when you want to build a complete index rather than query one firm at a time.
For RAG, the structured facts (name, number, status, officers, control) slot neatly into metadata and make excellent filters and grounded answers about UK businesses: verifying a company exists, checking who runs it, or tracing directorship and ownership links across companies. The Open Government Licence v3.0 lets you reuse all of it, including commercially, as long as you attribute the source.
The main caveat is depth versus breadth. The bulk snapshot carries only the basic fields, so the detailed content, full accounts and individual filings, comes per company through the API, and many of those documents are `PDF` or iXBRL that you will need to parse before they are retrievable text. Data is also self-filed by companies, so it is only as accurate and current as each company's last submission.
It sits naturally alongside SEC EDGAR for US corporate filings and Hansard or other UK government data when you are building a broader UK public-record assistant.
Related sources
data.europa.eu
The European Data Portal, a single point of access to open data from 36 European countries and the EU institutions. Over 1.6 million datasets span every policy area, giving RAG systems broad, multilingual coverage of official European data.
data.gov
The US federal government's open data portal, gathering more than 300,000 datasets from federal agencies. Coverage spans agriculture, climate, education, energy, finance, health, and public safety, making it a broad source of official US data for RAG systems.
data.gov.uk
The UK Government's central catalogue of open public sector data, run by the Government Digital Service. More than 47,000 datasets cover health, housing, transport, demographics, and the environment, giving RAG systems a trusted base of UK facts, figures, and official records.
data.police.uk
Crime and policing data published on behalf of police forces in England and Wales, Northern Ireland, and the British Transport Police, gathered in one central place. Covers street-level crime, outcomes, and stop-and-search records.