Lead enrichment from company websites

1. 01Start from domains

   The input is a public hostname, with or without https://. Domains usually come from the CRM, a sign-up form, or another API: the Google Maps API returns each local business's website, and the Product Hunt API returns each launch's website. IP addresses and localhost are rejected before any fetch.

   GET /v1/contacts?domain=stripe.com · pages

   [
     {
       "url": "https://stripe.com/",
       "type": "home"
     },
     {
       "url": "https://stripe.com/about",
       "type": "about"
     },
     {
       "url": "https://stripe.com/privacy",
       "type": "privacy"
     }
   ]

2. 02Look up what the company published

   One call reads the homepage and up to seven linked pages on the same site, choosing contact, about, imprint, team, support, and privacy pages first (pages sets the total; default 5, max 8). The price is one request whatever pages you pass. It returns company details from the site's metadata, then every email, phone, social profile, address, and named person it found, each with sources: the URLs it came from.

   GET /v1/contacts?domain=stripe.com · company

   {
     "name": "Stripe",
     "legal_name": "Stripe, LLC",
     "description": "Stripe powers online and in-person payment processing and financial solutions for businesses of all sizes.",
     "logo": "https://images.stripeassets.com/fzn2n1nzq965/1hgcBNd12BfT9VLgbId7By/01d91920114b124fb4cf6d448f9f06eb/favicon.svg",
     "country": "US",
     "language": "en-US"
   }

3. 03Pick the fields worth keeping

   emails[] marks each address generic (a role inbox such as sales@) or personal, and domain_match says whether it is on the queried domain; match_email_domain=true drops the rest. Phones come normalized to E.164 with the country. people[] lists names the site published, often without a job title or email, so treat it as a lead to research rather than a contact.

   GET /v1/contacts?domain=stripe.com · emails

   [
     {
       "value": "privacy@stripe.com",
       "type": "generic",
       "domain_match": true,
       "sources": [
         "https://stripe.com/privacy"
       ]
     }
   ]

# Enrich a list of company domains with the contacts they publish.
# In: domains.txt, one domain per line. Out: contacts.csv, one row per domain.
import csv, os, requests

API = "https://api.clair.im"
HEADERS = {"Authorization": f"Bearer {os.environ['CLAIR_API_KEY']}"}

def lookup(domain):
    r = requests.get(f"{API}/v1/contacts", headers=HEADERS, timeout=90,
                     params={"domain": domain, "pages": 5})
    r.raise_for_status()
    return r.json()

def best_email(emails):
    # An address on the company's own domain first, role inboxes before people.
    ranked = sorted(emails, key=lambda e: (not e["domain_match"], e["type"] != "generic"))
    return ranked[0]["value"] if ranked else None

with open("domains.txt") as f:
    domains = [line.strip() for line in f if line.strip()]

with open("contacts.csv", "w", newline="") as f:
    out = csv.writer(f)
    out.writerow(["domain", "company", "email", "email_source", "phone",
                  "linkedin", "country"])
    for domain in domains:
        try:
            c = lookup(domain)
        except requests.HTTPError as e:
            print(domain, e.response.status_code, e.response.text[:200])
            continue
        email = best_email(c["emails"])
        source = next((e["sources"][0] for e in c["emails"] if e["value"] == email), None)
        linkedin = next((s["url"] for s in c["social"] if s["network"] == "linkedin"), None)
        company = c["company"] or {}
        out.writerow([domain, company.get("name"), email, source,
                      c["phones"][0]["e164"] if c["phones"] else None,
                      linkedin, company.get("country")])

Set CLAIR_API_KEY to a key subscribed to the Website Contacts API. Each call is one request against that API's monthly quota.

• Only what the company published. There is no inference of email patterns (first.last@), no third-party database, and no verification that an inbox accepts mail. Run a deliverability check before sending.
• The lookup follows links from the homepage. It does not walk the sitemap or log in, so a contact page not linked from the homepage can be missed.
• Personal data rules still apply. An address a company published is not consent to email the person behind it; check the rules where you and the recipient are, such as GDPR in the EU and UK.
• Many companies publish only a form, not an email. An empty emails[] is a real answer, and the phone or LinkedIn profile is often the better contact.
• Use results in line with each site's terms, and do not resell them as a contact directory.