Lead-Gen Pipeline Pro

Feed it a list of business websites or domains. Get back a clean lead list: deliverable, scored contacts in one run. It crawls each site for emails, phones, and social profiles, then validates every email inline (DNS/MX, disposable, role, placeholder filtering) and surfaces the deliverable ones first. Validation is on by default, not a paid add-on you bolt on afterward.

The point: most contact scrapers hand you info@ regex soup that you then have to clean and verify with a second tool. This does the find-and-verify in one pass, so what comes out is already usable.

What a naive scraper gives you vs this

Run a plain regex scraper on a real business site and you get a pile like this, all mixed together:

billing@example.com        placeholder in the page template
jane.doe@example.com       demo address
logo@2x.png                an image filename that looks like an email
git@github.com             from a code snippet on the page
key@o0.ingest.sentry.io    a Sentry error-tracking DSN, not a person
hello@therealcompany.com   the one address you actually wanted

This pipeline strips the noise (placeholders, demo addresses, dev keys, third-party addresses smuggled in mailto params, image filenames) and grades what is left, so the real deliverable contact comes out on top with a score.

Output

One row per site, with the deliverable contacts surfaced first:

{
  "domain": "basecamp.com",
  "bestEmail": "jason@basecamp.com",
  "bestVerdict": "valid",
  "bestScore": 100,
  "deliverableCount": 1,
  "deliverableEmails": ["jason@basecamp.com"],
  "bestPhone": null,
  "emails": [{"email": "jason@basecamp.com", "verdict": "valid", "score": 100, "roleBased": false, "disposable": false, "freeProvider": false}],
  "phones": [],
  "socials": {"twitter": ["https://twitter.com/basecamp"]},
  "pagesScanned": 3,
  "status": "ok"
}

The default table view is a lead list: domain, best email, verdict, score, deliverable count, phone. Sort by score and start at the top.

• bestVerdict / each email's verdict is one of valid (deliverable), risky, unknown (we could not check the DNS right now), disposable, invalid. Only valid counts toward deliverableCount.
• score is 0-100 deliverability confidence (MX present, not disposable, not a role inbox).
• Role inboxes like info@ or sales@ are flagged with roleBased: true and can still verify as deliverable, since they do accept mail. They just score lower because they are weaker for personal outreach. Filter on roleBased if you only want personal addresses.

Input

{ "startUrls": ["https://apify.com", "basecamp.com"] }

• startUrls (required): business website URLs or domains, one per line.
• maxPagesPerSite: home page plus this many contact/about/team pages (default 4).
• followContactPages: auto-find and scan contact/about pages (default on).
• maxEmailsPerSite, maxPhonesPerSite: caps, keep the highest-scoring.
• proxyConfiguration: optional, for sites that block datacenter IPs.

Note: it takes a list of sites to scan. It does not yet discover sites from a keyword or location (that is on the roadmap). If you already have a list of company domains, this turns them into verified leads.

Billing (tiered, you pay more only as you get more)

• site-scanned: one per site successfully fetched (the crawl work). Cheap.
• contact-found: one per site that yielded at least one usable contact.
• verified-deliverable: one per email returned with a deliverable verdict. The premium tier, the part you actually want.

Sites we cannot fetch (DNS failure, timeout, hard block) are never billed. Email DNS failures come back as unknown and are not counted as deliverable, so you are never charged for a guess.

Honest limits

Server-rendered HTML only (no JavaScript execution). Phones come only from tel: links, because text-regex phone extraction is wrong more often than right. There is no SMTP handshake (those are blocked from shared IPs and unreliable), so deliverability is DNS-tier confidence: it proves the domain accepts mail, not that a specific person's mailbox exists. A real but fictional demo address on a real domain can still pass. Sites behind heavy bot protection may need the proxy option.