Data sources

Every figure on this site comes from Companies House — the UK's public company register run by the Department for Business and Trade. Nothing is sourced from private datasets, paid feeds, or third-party brokers. Below is exactly what we use and how often we refresh it.

What we ingest

Free Company Data Product (bulk monthly)

Companies House publishes a full snapshot of the live register on the first of each month at download.companieshouse.gov.uk. It contains every active UK company's basic record: name, number, type, status, incorporation date, registered office address, accounts and confirmation-statement filing dates, and up to four SIC industry codes. We load the whole snapshot — approximately 5.7 million companies — into the Company and CompanySIC tables on the 1st of each month.

Important: this product contains live-register companies only. Dissolved companies are removed when they're struck off. For dissolved companies we fall back to the live API (below).

People with Significant Control (PSC) snapshot (bulk, daily)

A daily JSONL file at download.companieshouse.gov.uk/en_pscdata.html listing every UK PSC: name, kind (individual or corporate), nature of control, nationality, year/month of birth (no day, by law), country of residence. We hold approximately 46 million PSC rows. Loaded incrementally as new daily files publish.

iXBRL accounts data (bulk, daily)

Daily zip files at download.companieshouse.gov.uk/en_accountsdata.html containing every accounts filing submitted that day in machine-readable iXBRL format. We parse each filing and extract roughly 20 standard FRS-102 concepts (turnover, net assets, cash, etc.). At present we hold accounts figures for around 780,000 UK companies and grow this daily.

Older accounts filings (typically pre-2020 for established companies) are PDF-only and not currently extractable without OCR. P&L figures (turnover, profit) are only publicly disclosed by medium and large companies — small companies file abridged or micro-entity accounts under UK law, where the P&L is not in the public record.

Companies House live REST API (on-demand, free tier)

For pages where bulk data is missing or insufficient, we live-fetch from api.company-information.service.gov.uk:

Our application is registered with Companies House Developer Hub as "Company Analyst" under the standard free tier (600 requests / 5 minutes).

What we do with the data

We display it. We compute basic comparisons — for example, sector medians and quartiles within size buckets, so you can see how a company compares to its peers. We use the Anthropic Claude API to generate a short narrative summary of a company's filed figures, with strict instructions never to invent numbers not present in the source data. The narrative is cached per company so the same analysis is shown to every visitor — Anthropic doesn't see your visit, only the underlying public data.

What we don't do

Accuracy and corrections

We aim to mirror Companies House faithfully but data is only as current as the most recent refresh. If something is wrong on a company's filing at Companies House, it'll be wrong here until Companies House correct it and we re-ingest. To request a correction to the underlying record, contact Companies House directly. To request a redaction on our display while the underlying record is updated, email hello@companyanalyst.co.uk.

Companies House data is published under the Open Government Licence v3.0. Required attribution: "Contains public sector information licensed under the Open Government Licence v3.0."