Contact Finder Pro — MX-Verified Emails, Validated Socials, Role Detection

Premium contact-info extractor. Built for B2B teams who need contact data they can actually use — not regex dumps full of dead emails and broken social links. MX-verified emails — every email's domain is checked for live MX records. Junk gets dropped, not shipped. Role classification — sales@, careers@, support@, legal@, press@, personal automatically tagged so you can target the right inbox. Phone E.164 normalization — international format ready for dialer/CRM import. Social profile HEAD validation — Facebook, Instagram, Twitter/X, LinkedIn, YouTube, TikTok, GitHub URLs verified to actually resolve before they're emitted. Multi-page crawl per domain — homepage + /contact, /about, /team, /careers, /privacy (where DPO emails live). Firecrawl anti-bot fallback — when direct fetch fails on Cloudflare/Akamai sites, retry through full browser rendering.

Why Pro

The cheap contact-info scrapers on Apify Store ship raw extraction. They regex any something@something.com they find, dump the result, and call it done. The most popular alternative, vdrmota/contact-info-scraper (3.59 ★, 79 reviews), has 99% reliability — the complaint isn't that it fails. The complaint is the output is unusable: bounced emails, broken social URLs, no role context, raw phones in 5 different formats.

Pro is built around the validation gap:

Layer	Pro	Incumbent
Email MX verification (drops dead domains)	Yes	Mostly no
Email role classification	sales / support / careers / legal / press / billing / personal / info	Raw email only
Phone normalization (E.164)	International format	As-found
Social profile validation (HEAD 200 check)	Yes	Raw URL only
Multi-path crawl per domain	6 paths, single price	Configurable but billed per page
Anti-bot fallback (Firecrawl)	Built-in	Manual config
Per-record quality score	0.0-1.0	None
JSON-LD Organization schema parsing	Yes	No
Pricing model	PPE — pay per delivered VALUE	PPE — pay per page scraped

What You Get

For each input URL, one structured record:

Core fields

• inputUrl — what you submitted
• domain — extracted domain (lowercased, no www.)
• companyName — pulled from JSON-LD Organization, og:site_name, og:title, or <title>
• pagesFetched — how many candidate pages were actually retrieved
• firecrawlUsed — true if Firecrawl rescue was triggered
• extractedAt — ISO 8601 UTC timestamp
• qualityScore — 0.0-1.0 weighted completeness

Emails

"emails": [
  { "email": "sales@stripe.com", "role": "sales", "verified": true },
  { "email": "press@stripe.com", "role": "press", "verified": true },
  { "email": "info@stripe.com",  "role": "info",  "verified": true }
]

Roles: sales, support, careers, legal, press, billing, info, personal.

Phones

"phones": [
  { "raw": "+1 (415) 555-2671", "e164": "+14155552671" }
]

Numbers that fail E.164 normalization but contain at least 10 digits are still emitted with e164: null so you can normalize them yourself.

Socials

"socials": [
  { "platform": "twitter",  "url": "https://twitter.com/stripe",  "validated": true  },
  { "platform": "linkedin", "url": "https://linkedin.com/company/stripe", "validated": true },
  { "platform": "github",   "url": "https://github.com/stripe",   "validated": false }
]

Platforms detected: facebook, instagram, twitter (X), linkedin, youtube, tiktok, github. validated: true means the profile URL responded HTTP 200 to a HEAD request.

Quality SLA

Pro enforces a contract on every run before publishing the dataset:

Metric	Target	What it means
Min items returned	1	If a query returns nothing, the run fails loudly instead of silently shipping an empty dataset.
Min contact rate	50%	At least half of input URLs must return at least one email/phone/social.
Min average quality score	0.30	Average completeness across the dataset.

If the dataset misses these targets, the run is flagged so you can re-run on a different input/proxy instead of paying for junk.

Input Examples

Quick test (3 domains)

{
  "urls": [
    "https://www.intercom.com",
    "https://stripe.com",
    "https://shopify.com"
  ]
}

CRM enrichment (large list, full validation)

{
  "urls": ["acme.com", "globex.io", "initech.co.uk"],
  "verifyEmailMx": true,
  "firecrawlFallback": true,
  "validateSocialProfiles": true,
  "maxPagesPerDomain": 8,
  "maxConcurrency": 5
}

Cost-controlled bulk run (skip social validation)

{
  "urls": ["..."],
  "validateSocialProfiles": false,
  "firecrawlFallback": false,
  "maxPagesPerDomain": 3
}

Region-specific (UK phones)

{
  "urls": ["https://www.john-lewis.co.uk", "https://tesco.com"],
  "countryHint": "GB"
}

Output Example

{
  "inputUrl": "https://www.intercom.com",
  "domain": "intercom.com",
  "companyName": "Intercom — The only helpdesk designed for AI-first customer service",
  "emails": [
    { "email": "team@intercom.com", "role": "personal", "verified": true }
  ],
  "phones": [{ "raw": "+1 (415) 660-9690", "e164": "+14156609690" }],
  "socials": [
    {
      "platform": "twitter",
      "url": "https://twitter.com/intercom",
      "validated": true
    },
    {
      "platform": "linkedin",
      "url": "https://www.linkedin.com/company/intercom",
      "validated": true
    },
    {
      "platform": "facebook",
      "url": "https://www.facebook.com/intercom",
      "validated": true
    },
    {
      "platform": "youtube",
      "url": "https://www.youtube.com/c/intercom",
      "validated": true
    },
    {
      "platform": "github",
      "url": "https://github.com/intercom",
      "validated": false
    }
  ],
  "pagesFetched": 5,
  "firecrawlUsed": false,
  "extractedAt": "2026-05-06T22:46:57Z",
  "qualityScore": 0.9
}

Pricing — Pay Per Event (effective 2026-05-20)

Charged only when delivered value matches:

Event	Price	When charged
actor-start	$0.05	Once per run
url-processed	$0.02	Per input URL with at least one contact field returned
email-verified	$0.05	Per email confirmed deliverable via MX
social-validated	$0.02	Per social profile that responded HTTP 200

Realistic run costs

Scenario	Breakdown	Total
100 domains, 30 returned 1+ verified email + 3 socials	$0.05 + 100 × $0.02 + 30 × $0.05 + 90 × $0.02	$5.35
50 domains, full enrichment, 40 hits	$0.05 + 50 × $0.02 + 40 × $0.05 + 80 × $0.02	$4.65
1,000 domains bulk, no enrichment	$0.05 + 1000 × $0.02	$20.05

You pay for delivered VALUE, not for infrastructure. If a domain returns nothing, no url-processed charge fires.

Best Use Cases

1. B2B sales prospecting — feed your TAM domain list, get sales@/info@ contacts you can mail.
2. CRM enrichment — append role-tagged emails, normalized phones, validated socials to existing accounts.
3. Investor research — extract press@, founder personal emails, and verified LinkedIn/Twitter profiles from portfolio companies.
4. Recruitment — pull careers@/jobs@ contacts from target companies for outbound recruiter outreach.
5. Compliance / DPO discovery — many privacy@ and dpo@ emails live on /legal/privacy pages — Pro's multi-path crawl finds them.
6. PR / media outreach — press@ and media@ emails pre-classified, ready for distribution lists.
7. Competitor intel — quickly gather public contact surface from competitor sites with quality score 0.6+.

Programmatic Usage

Python

from apify_client import ApifyClient

client = ApifyClient("YOUR_TOKEN")
run = client.actor("tugelbay/contact-finder-pro").call(run_input={
    "urls": ["intercom.com", "stripe.com", "shopify.com"],
    "verifyEmailMx": True,
    "firecrawlFallback": True,
    "validateSocialProfiles": True,
})

for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    sales_emails = [e["email"] for e in item["emails"]
                    if e["verified"] and e["role"] == "sales"]
    if sales_emails:
        print(f"{item['domain']:<25} -> {sales_emails[0]}")

JavaScript / TypeScript

import { ApifyClient } from "apify-client";
const client = new ApifyClient({ token: "YOUR_TOKEN" });

const run = await client.actor("tugelbay/contact-finder-pro").call({
  urls: ["acme.com", "globex.io"],
  verifyEmailMx: true,
});

const { items } = await client.dataset(run.defaultDatasetId).listItems();
const verified = items
  .flatMap((it) => it.emails.filter((e) => e.verified))
  .filter((e) => e.role === "sales");
console.log(`${verified.length} verified sales emails`);

LangChain tool

from langchain_core.tools import tool

@tool
def find_contacts(domain: str) -> dict:
    """Get verified contact info (emails by role, phones, socials) for a company domain."""
    from apify_client import ApifyClient
    client = ApifyClient("YOUR_TOKEN")
    run = client.actor("tugelbay/contact-finder-pro").call(run_input={
        "urls": [domain],
        "verifyEmailMx": True,
        "firecrawlFallback": True,
    })
    items = list(client.dataset(run["defaultDatasetId"]).iterate_items())
    return items[0] if items else {}

MCP (Streamable HTTP)

npx mcpc connect mcp.apify.com
npx mcpc call tugelbay/contact-finder-pro --json '{"urls":["stripe.com"]}'

FAQ

Q: How is "MX-verified" different from full SMTP probe? A: MX verification confirms the domain is set up to receive mail at all — it catches ~80% of dead/junk addresses (typos, parked domains, abandoned sites) without the false-positive rate of SMTP probing (which many mail servers refuse). For deeper validation pipe email through NeverBounce or ZeroBounce after this run.

Q: Why don't most modern SaaS sites return any emails? A: They moved to contact forms behind /contact-sales. There is no email on the page to extract. Pro will still return companyName, validated socials, and any phones. For raw email harvesting at SaaS companies, follow up with a tool that extracts founder/employee emails from LinkedIn (separate actor).

Q: Will Firecrawl be charged when my run uses it? A: Firecrawl calls happen on the actor's compute cost, not yours. The premium PPE pricing already accounts for fallback usage. You don't pay separately.

Q: What if a domain returns nothing? A: No url-processed charge is fired (you only pay for VALUE). The actor logs no contacts found for X and moves on.

Q: Does it follow redirects? A: Yes — httpx follows redirects, and the final URL's domain is what gets emitted in the domain field.

Q: International support? A: All extraction works globally. For phone normalization, set countryHint to the most-common country code in your input list (GB, DE, FR, etc.). US is the default.

Q: How does this differ from official Hunter.io / Apollo / ZoomInfo APIs? A: Those services charge per-lead with monthly minimums ($49+ Hunter, $99+ Apollo). Pro is PPE — pay only when you actually receive a verified contact. For occasional prospecting, dramatically cheaper.

Troubleshooting

Empty dataset across all URLs

• Check that urls is set; the actor falls back to a demo URL and warns if not.
• Some domains (LinkedIn, Twitter accounts directly) are heavily rate-limited — try variants.
• Disable firecrawlFallback to rule out timeouts at the fallback layer.

Many emails returned but verified: false

• Domain might have temporary DNS issues — retry the run.
• The MX lookup falls back to 8.8.8.8 / 1.1.1.1 automatically; if your domain has authoritative-only DNS, verification will be inconclusive.

Phones look like CSS color codes

• Pro filters phones to E.164-normalized OR 10+-digit. If you still see noise, lower maxPagesPerDomain to 3 (skips footer/legal pages where SPA color tokens live).

Slow runs

• Set validateSocialProfiles: false to skip the per-profile HEAD checks (saves about 3s per social URL).
• Set firecrawlFallback: false if you don't need the anti-bot rescue.
• Drop maxPagesPerDomain to 3 (homepage + /contact + /about only).

Limitations

• Email harvesting depends on the site exposing emails on its public pages. Modern SaaS rarely does — Pro will return companyName + validated socials + phones even when emails are zero.
• MX verification is necessary but not sufficient. A domain with MX records can still have a non-existent mailbox. Pipe email through SMTP-probe service for deeper validation.
• HEAD validation may miss profiles that block HEAD but allow GET. Pro retries with GET on 403/405, but a few platforms (Instagram for non-public profiles) return 200 even for nonexistent users.
• Phone normalization is best-effort. If phonenumbers library cannot parse the number, the raw is still emitted (with e164: null) when 10+ digits.
• Crawl is shallow by design — only the listed contact paths, not full-site crawl. For deep crawl, use a separate Apify crawler upstream.

Changelog

• 0.1 (2026-05-06) — Initial release. Multi-path crawl, MX verification, role classification, E.164 phone normalization, social HEAD validation, Firecrawl fallback, JSON-LD parser, quality score. Premium PPE pricing scheduled for 2026-05-20.