Company Dossier from a Domain (BE / NL / FR / LU / DE) — contact route, VAT and company number in one run

Give a list of company websites. Each site is read once, and you get one row per company with:

• the best public contact route (email, phone or contact form) with the page it came from;
• the company number and VAT number published on the site, checked with the official checksum rules and, optionally, with the European Commission's VIES service;
• optionally, whether that number is registered for Peppol e-invoicing, checked with the Peppol network's own lookup;
• optionally, booking and quote-request routes (online appointment, request a quote).

Use it to: enrich a lead or supplier list in one step, fill in VAT numbers for invoicing, check that a website names a valid company, or clean a list from a business-listing or Google Maps actor (pick its dataset as input).

This actor combines two engines that are also available on their own: the Contact Profile actor and the Company Legal ID Finder. One crawl is shared, so a dossier costs fewer page requests than running both.

What you get per website

Field	What it means
status	complete (contact and identity), contact_only, identity_only, nothing_found, unreachable, redirected_to_other_domain, failed, invalid_input
contact	Contact route of the Contact Profile engine: chosen email (role or named address on the company's own domain), phone, form, sources and evidence
company_identity	Country, main identifier, all identifiers of the site owner, VIES result, name match, confidence and reasons
identifiers / third_party_identifiers	Every number with its source page; numbers of hosting providers, agencies and payment providers are kept apart
peppol	Only when switched on: registered (with participant ID and the host name of the service provider's SMP server), not_registered, source_unavailable or not_applicable
conversion_routes	Only when switched on: proven booking/quote routes and candidate order/price routes, with evidence

How it works

• Reads only the company's own website: the same domain, robots.txt respected, at most 5 pages. Contact and legal-notice pages are read first. With conversionRoutes on, the page behind a booking or quote button is also opened (at most 8 per website, also when it is on a booking platform), through the same protected transport.
• No forms are sent, no logins, no mailbox checks; listing and review platforms are not scraped.
• VIES requests are spaced out and cached per run.
• Peppol: the identifiers found on the website are looked up in the Peppol SML, the central register of the e-invoicing network, through public DNS. No directory search, no account. At most 3 lookups per company. A busy or unavailable service is reported as source_unavailable, never as "invalid" or "not listed".

Pricing

Pay per event (USD). Every run also carries Apify's standard start event (apify-actor-start): $0.00005 per run, or 5 cents per 1,000 runs.

Event	When	Price
website-checked	every website that was checked (failures on our side are free)	$0.003
company-dossier	an accepted contact route or a valid company identity was found	$0.03

• VIES and Peppol checks, the booking/quote routes and the contact options (channelScope: all for social profiles, WhatsApp and the booking link; emailScope: all for freemail and department addresses) are included.
• Rows are stored before they are charged; websites the run had no time for are left for a resumed run.
• Example: 1,000 websites with a dossier for 600 of them cost about $3 + $18 = $21.

Limits — please read

• Everything is what the website itself publishes. An email or phone is a published route, not a verified mailbox or a consent to be contacted. Check the rules that apply to your outreach (for example GDPR and national direct-marketing rules).
• A number on a website is a claim of that website. VIES confirms that a VAT number is registered. It does not prove that the website belongs to that company; name match and confidence are signals, not proof.
• Peppol not_registered means no registration was found for the identifiers on the website. The company may receive invoices under another identifier.
• Many small businesses publish no company number on their website. Expect an identity for a minority of websites (see below).
• PDF documents are not read in this version.

Measured results of the engines

The combined dossier has not been blind-tested as one product. The two engines were, separately:

• Company identity (October 2026, blind test on 400 unseen websites from BE, NL, FR, LU and DE): 103 websites gave an identity; for 99 of the 103 the main identifier was the right company's number, copied exactly with the right type. 108 of the 400 websites could not be reached.
• Contact route (preregistered comparison on 300 SME websites, October 2026): 152 companies got a correct address with all email categories, at 97.4% precision. A re-score of the same saved scans for the default categories gave 139 companies at 97.9% precision; that re-score is not a separate independent test.
• Booking and quote routes (blind test on 660 new SME websites in BE, NL and FR): 48 of 50 proven routes were correct. Order and price routes are returned as candidates and are not covered by that figure.

These are results on test samples, not a guarantee for your list. Try a sample of your own websites first.

Data sources and terms

• The websites you provide, read according to their robots.txt.
• VIES, the VAT Information Exchange System of the European Commission (public service, no key needed).
• Peppol SML (Service Metadata Locator), the public DNS register of the Peppol network, run by OpenPeppol under participant.sml.prod.tech.peppol.org (only when peppolCheck is on).