Company Enrichment from Domain

Company enrichment and domain enrichment from the company's own website. Give it a list of company domains (website to company data). For each one you get a single row: the company name, a short description, role email addresses, phone numbers, social links, the schema.org postal address, the website's CMS, ecommerce platform and analytics tags, the email provider, SPF and DMARC records, the SSL certificate expiry date and the domain registration date.

• Where the data comes from: only the company's own website and public DNS and RDAP (the registries' successor to WHOIS). Nothing is bought, guessed or taken from a third-party database.
• What it does not have: no employee counts, no revenue figures and no other estimates. It is not a people database: emails are role addresses (info@, sales@...) on the company's own domain or a free-mail inbox it publishes.
• Bot-walled sites return status only: a site that blocks automated readers (HTTP 403/503, or a bot-check or queue page) gets its HTTP status and the error bot wall or http 403, plus DNS facts when includeDns is on. Such rows are free.

Pricing

$0.006 per enriched domain (event domain-enriched). A row is charged only when the company's own website was read: it answered with a status below 400, was not a bot wall and gave at least one fact. DNS and RDAP facts (email provider, SPF, DMARC, registration date) are included at no extra charge. Free rows: invalid domains, and sites that are unreachable, robots-blocked, blocked (403/503) or bot-walled, even when their DNS facts are returned. Every row has a charged column that says whether it was billed. A domain listed twice is profiled and charged once.

Input

Input	Default	What it does
domains	5 sample domains (prefill)	Required. Domains or website URLs, up to 500 per run. example.com, www.example.com and https://example.com/contact all mean example.com. A domain listed twice is profiled once. An entry that is not a domain gets a row with error invalid domain.
maxPagesPerDomain	3	Pages read on each site, homepage included (1 to 5). 1 reads the homepage only. 2 adds the contact page. 3 adds the about page. 4 and 5 add linked team, location, company or imprint pages.
includeDns	true	Look up the email provider, SPF, DMARC and the registration date. Turn it off for a faster, website-only profile.

{ "domains": ["mailchimp.com", "zapier.com", "buffer.com", "squarespace.com"], "maxPagesPerDomain": 3, "includeDns": true }

Output

You get exactly one dataset row per unique input domain, in input order, with these keys in this order:

Field	Meaning
domain	The normalized domain (example.com). For an invalid entry, the text you entered.
finalUrl	The homepage URL after redirects.
httpStatus	The HTTP status of that homepage.
companyName	The company's name. Taken from schema.org first, then og:site_name, then the page <title> (see below).
companyNameSource	Where the name came from: schema.org, og:site_name, title, or null.
description	The homepage meta description. If there is none, og:description. If there is none, the schema.org description. At most 300 characters.
emails	Role addresses only, such as info@, sales@, support@ and hello@, from every page read and from schema.org email, kept only when they are on the company's own domain (or its redirect target) or a common free-mail inbox. Template placeholders (email@address.com, you@domain.com...) and addresses on other companies' domains (a web agency, a sister business) are dropped. Addresses that look like a person's name (jane.smith@...) are dropped on purpose. Sorted, at most 10.
phones	Phone numbers from tel: links, North American numbers in the visible text, and schema.org telephone. Template placeholder numbers are dropped. Sorted, at most 6.
socialLinks	{network: url} for Facebook, Instagram, LinkedIn, X, YouTube, TikTok, Yelp, Pinterest and Nextdoor. Links on the pages come first, then schema.org sameAs. The first link found for a network is kept. Share buttons are ignored.
postalAddress	{streetAddress, addressLocality, addressRegion, postalCode, addressCountry} from a schema.org PostalAddress, written inline or referenced by @id on the same page. null when the site publishes no structured address. The actor never guesses an address from page text.
cms	The site builder or CMS, such as WordPress, Wix, Squarespace or Webflow. Detected from the homepage markup, response headers and the generator tag.
ecommercePlatform	Shopify, WooCommerce, BigCommerce or Magento when detected.
analytics	Analytics and advertising tags found on the homepage, such as Google Analytics, Google Tag Manager, Meta Pixel and Microsoft Clarity. Sorted.
emailProvider	Who handles the domain's email, based on its lowest-preference MX host. Values include Google Workspace, Microsoft 365, Zoho, GoDaddy, Proofpoint, Mimecast, Barracuda, Rackspace, Yahoo, iCloud, Fastmail, Proton, Namecheap, IONOS, Hostinger and Cloudflare Email Routing. Any other host gives its registrable domain. A null MX record gives none (null MX). No MX record gives null.
spf	The domain's SPF record (v=spf1 ...), or null.
dmarc	The domain's DMARC record (v=DMARC1; ...), or null.
sslExpiry	When the website's TLS certificate expires (its notAfter), in ISO-8601 UTC. null when the site is not served over https or the certificate does not verify.
domainCreated	The domain's registration date from RDAP. For a subdomain, the registered parent's date.
pagesChecked	The pages that were read, homepage first.
error	null on success. Otherwise a short reason: invalid domain, bot wall (a bot-check or queue page answered instead of the site; status only), robots-disallowed (also an unreadable robots.txt or a crawl delay that is too long), blocked: non-public address, host asked us to back off, too many redirects, bad redirect, unreachable: <reason> (for example unreachable: gaierror when the domain does not resolve), http 404 (or another status), not html, certificate error, dns lookups failed, dns failed: <reason> or unexpected failure: <reason>.
charged	true when this row was billed one domain-enriched event (see Pricing).

A row with an error still carries every fact that could be read. For example, a site with a broken certificate is still read, its facts are filled in, and the row says certificate error. DNS facts are filled in even when the website is down.

How the company name is chosen

1. schema.org JSON-LD: an Organization, LocalBusiness or one of their common subtypes, such as Store, Restaurant, Dentist, Physician, Attorney, Plumber, HomeAndConstructionBusiness, AutoDealer or RealEstateAgent. @graph, arrays and lists of types are all read. The names of WebSite, WebPage, Product and Person entries are never used. schema.org microdata (an organization itemscope with itemprop="name") counts as schema.org too, unless the organization is a property of another item, such as a Product's manufacturer or an Offer's seller.
2. The og:site_name meta tag.
3. The homepage <title>. If the title starts with a generic segment such as "Home" or "Welcome" followed by a separator (|, –, —, -, :), the next segment is used. Otherwise the whole title is used, so this fallback can include a tagline.

Which pages are read

The actor reads the homepage first. It tries https://domain/, then http://domain/, then https://www.domain/. Then, within maxPagesPerDomain, it reads more pages on the same site:

1. The linked contact page and the linked about page.
2. If either is not linked, /contact or /about instead.
3. Only when maxPagesPerDomain is 4 or 5: linked team, location, company, imprint or impressum pages.

No URL is requested twice. Redirects are followed one hop at a time, so a redirect to a page that was already read is not requested again. A redirect that only adds or removes the trailing slash (/contact to /contact/) is followed. Pages that are not HTML, or that return an error status, are skipped. Page text is never included in the output.

How it behaves on other people's sites

• Every request identifies itself as WeioBot.
• It obeys robots.txt for every host and every redirect hop. If robots.txt refuses the homepage, the row says robots-disallowed and nothing else on that site is fetched. The actor never tries another route around it.
• Requests are paced to about one per second per host (slower if robots.txt asks for a crawl delay). The pacing is shared by every domain in the run.
• When a host answers 429 or 503, the actor waits as long as the host asks (Retry-After) and then retries. If the host asks for a pause longer than 30 seconds, the actor does not wait. The page is skipped instead.
• Only public internet addresses are contacted.
• For sslExpiry, the actor opens one verified TLS connection to the site and reads the certificate. It sends no HTTP request on that connection.
• DNS facts come from public DNS-over-HTTPS (Cloudflare, with Google as the fallback). Registration dates come from RDAP, the registries' official successor to WHOIS, found through the IANA bootstrap file. Those services are reached the same way: WeioBot identity, their robots.txt obeyed, paced, with Retry-After backoff on rate limits.

Limits

• This actor does not run JavaScript. A site that renders all of its content with JavaScript may yield only its title.

• Only role email addresses are returned. If a site lists only personal addresses, emails is empty.

• A row is charged when the company's own website gave at least one fact (see Pricing), whether or not the site publishes a role email, phone number or social link.

• Phone numbers in the visible text are recognized only in North American format. tel: links and schema.org numbers in any format are kept.

• Technology detection looks at the homepage only, using a compact rule set, not a full signature database.

• With DNS on, each domain needs about 6 to 7 DNS-over-HTTPS lookups. Those lookups share one DNS host that is paced to one request per second, so a large list takes about 6 to 7 seconds per domain. The website part runs up to 5 domains at once.

• domainCreated is null when the registry has no RDAP service or does not publish a registration event.

• The default run timeout is 1 hour, which fits about 500 domains with DNS on (about 7 seconds each). For longer lists, split them or raise the timeout in the run options. If a run stops early you are charged only for the rows it delivered.

• A bot-check page is recognised by its title or challenge markers (Cloudflare, PerimeterX, DataDome, Incapsula and similar). A new kind of wall may not be recognised; its row then has the page's own title as companyName.

Built and maintained by Weio, Inc. The actor was written and is operated by AI agents (Anthropic Claude and OpenAI Codex), with a person accountable at Weio.