Email verification works by running a series of checks in order: syntax, domain and MX records, and then (where the server allows it) an SMTP-level mailbox check. Each stage rules out more bad addresses, and none of them can guarantee delivery on its own. This guide shows how to build the first stages in Python and explains where the limits are.
What does email verification actually check?
- Syntax: does the address have a valid local part and domain?
- Domain and MX lookup: does the domain exist and publish mail exchangers?
- Disposable and role detection: is it a throwaway domain, or something like
info@oradmin@? - SMTP probe: does the receiving server accept
RCPT TOfor this mailbox?
Many servers accept all recipients (catch-all) or throttle probes, so the SMTP result is often "unknown" and should be treated that way.
Step 1: syntax and MX check
import re import dns.resolver # pip install dnspython EMAIL_RE = re.compile(r"^[A-Za-z0-9._%+-]+@([A-Za-z0-9-]+\.)+[A-Za-z]{2,}$") def check_syntax(email: str) -> bool: return bool(EMAIL_RE.match(email)) def get_mx(domain: str) -> list[str]: try: answers = dns.resolver.resolve(domain, "MX", lifetime=5) return [str(r.exchange).rstrip(".") for r in sorted(answers, key=lambda r: r.preference)] except (dns.resolver.NXDOMAIN, dns.resolver.NoAnswer, dns.resolver.LifetimeTimeout): return [] Step 2: classify instead of returning true/false
ROLE = {"info", "admin", "support", "sales", "contact"} def verify(email: str) -> dict: if not check_syntax(email): return {"email": email, "status": "invalid", "reason": "syntax"} local, domain = email.rsplit("@", 1) mx = get_mx(domain) if not mx: return {"email": email, "status": "invalid", "reason": "no_mx"} status = "risky" if local.lower() in ROLE else "deliverable_domain" return {"email": email, "status": status, "mx": mx[0]} Step 3: a polite SMTP probe
import smtplib def smtp_probe(email: str, mx_host: str, sender: str = "verify@example.com") -> str: try: with smtplib.SMTP(mx_host, 25, timeout=10) as s: s.helo("example.com") s.mail(sender) code, _ = s.rcpt(email) if code == 250: return "accepted" if 500 <= code < 600: return "rejected" return "unknown" except (smtplib.SMTPException, OSError): return "unknown" Use your real sending domain, probe at a low rate, cache results per domain, and never treat "accepted" as proof the mailbox is real, because catch-all servers accept everything. Many cloud hosts also block outbound port 25.
Practical tips
- Run domain-level checks once per domain, not once per address.
- Keep "unknown" as its own status and send those to a small test batch.
- Remove hard bounces from your list right after each campaign.
- Verify at sign-up time to keep bad data out.
FAQ
Can verification guarantee an email will be delivered? No. It lowers bounce risk; inbox placement also depends on reputation and content.
What is a catch-all domain? One that accepts mail for any address, so the SMTP check cannot confirm a specific mailbox.
Is SMTP probing safe to run at scale? Only at low, rate-limited volumes from a properly configured sender; aggressive probing can get your IP blocked.
Should I build or buy? Building the syntax and MX stages is easy; reliable SMTP probing at volume needs infrastructure and monitoring.
Written by Muhammad Afzal, Python developer at INDUS Web Agency.

