# Changelog of UK Companies House Scraper — Companies, Officers, PSC & Filings (`teodor_banea/companies-house-scraper`) Actor

- **URL**: https://apify.com/teodor_banea/companies-house-scraper/changelog.md
- **Full Actor documentation**: https://apify.com/teodor_banea/companies-house-scraper.md

## Changelog

### 0.1.10 (2026-10-20)

- Charges: every page of the charges list is read (the list uses an older pager; companies with more than 25 charges
  stopped at page 1).
- Addresses: overseas addresses are split correctly — US state + ZIP ("Wy 82001"), countries glued to a ZIP
  ("95054 Usa", "Canada L3R 0H6"), foreign postcodes ("KY1-1104", "7735"), many more country names (Belize,
  Curacao…); a part with a digit is never taken as the town; repeated postcodes are dropped.
- Category-filtered filing histories are read to the newest 200 filings per company (was 500); the 0-row message
  names the categories; a warning when charges / category-filtered filings scan more than 1,000 companies.
- "N/A" / "NONE" / "-" in foreign registration fields become null; nationalities are written "British, Irish".
- Singular / plural in logs and status ("1 company"); progress status every 30 s with one update in flight.

### 0.1.1 – 0.1.9 (2026-10-20)

- Platform tuning: concurrency, queue order (a company's follow-up pages first, rows stream out), at most 3
  downloads at once, a bounded company backlog, heap cap (about 140 MB used in a 512 MB run), first-accounts dates,
  bare-premises address merge, region list without city names.

### 0.1.0 (2026-10-20)

- First version: the UK Companies House register from the public website, no API key.
- Company source: advanced search (name includes / excludes, registered office address, statuses, company types,
  subtype, SIC codes, incorporation and dissolution date ranges) or a list of company numbers / company page URLs.
- Data types: companies (full profile with split address, SIC descriptions, previous names, accounts and
  confirmation statement dates, overdue flags, charges / insolvency flags, officer counts), officers, persons with
  significant control (individuals, corporate, legal persons, statements, exemptions), filing history with PDF
  links and derived categories, charges.
- Searches above the site's 5,000-per-search download limit are split by incorporation date (newest first, only
  as far as the run needs); several SIC codes are searched one by one and de-duplicated.
- GDPR minimisation: month and year of birth only, no identity-verification narrative.
- Rate-limit aware: requests paced and spread over datacenter proxy sessions, back-off on the site's 403 block.
- Lean memory: the JS heap is capped at ~55 % of the run memory (about 140 MB used in a 512 MB run).