Email Format Finder by Domain

Pay per result: $1.796/1k on the Free plan down to $1.396/1k on Gold. Free Apify plan: up to 100 per run, as many runs as your credit covers. Misses and notices are free.

Email format finder by domain: give a list of company domains, get the work email pattern each company uses (first.last, flast, first and more), with the number of people behind it. No login, no cookies, no start fee. You pay only for a pattern you receive.

No login | Domains, URLs or emails in | One pattern per row | Thin evidence skipped, never guessed | Unknown domains = free | JSON/CSV export

Domain in, email format out. Cold outreach needs the shape of a company's addresses: is it first.last@, flast@ or first@? This actor reads up to 500 of a company's current people, looks at the addresses on record for them on that domain, and counts which name-based format each address follows. The format most people fit is the answer. It returns the pattern and the evidence, not a list of addresses. Where the evidence is thin, you get a free notice instead of a guess.

See all actors by b2bdata

---

💳 Pricing and plans

One event, email-pattern, charged once per pattern row delivered. Prices per 1,000 patterns (set on the Pricing tab):

Event	When it is charged	Free plan	Gold plan
email-pattern	A pattern row is delivered to your dataset	$1.796	$1.396

Bronze and Silver prices are on the Pricing tab. Higher Apify plans only lower the price. Empty runs, unknown domains, skipped domains, failed lookups, unreadable entries and notices are free, and there is no start fee.

Plan-credit math. The free plan's $5 monthly credit covers about 2,780 patterns at the Free price.

Success-only billing

You are charged only for a pattern that passes three floors: at least 5 people confirm it, at least 10 people have an address on the domain, and at least 40% of those people follow it. Everything else gets one free notice row and is listed in MISSES with its position and a reason: unresolved_domain, no_addresses, no_pattern, weak_evidence, low_confidence, shared_domain or unreadable_answer. A notice never shows the pattern. An entry that is not a domain, a repeat, or a lookup that fails is listed in MISSES and costs nothing.

The free plan

Free Apify accounts get up to 100 domains looked up per run, and as many runs as their Apify credit covers. Nothing is stored on our side. When a list is longer, the first 100 valid domains are looked up and one free notice row (reason: "free_plan_limit") says how many were not looked up; they are not charged. Run again with the rest, or use a paid Apify plan for no per-run limit.

Keep a run small while you test

limit caps the whole run (default and maximum 800). Try one domain with limit: 1, read the row, then scale up. A lookup takes about 3 to 4 seconds, so 800 domains take about 45 minutes.

---

⚡ Why this email pattern actor?

• Evidence, not a bare guess. Every row says how many people confirm the pattern, what share of the people with an address follow it, and which other formats appeared.
• It does not guess on thin data. One person fitting a format, or two, is not a company format. Those return a free notice.
• No addresses leave the actor. Sample addresses are masked by the data (j***.d**@acme.com). You get the shape, not a contact list. To find a person's address, combine the pattern with a name or use email-finder.
• Success-only, no start fee. Unknown domains, hosting platforms, thin patterns, bad entries and repeats are free.

---

🎯 Use cases: turn a domain column into email formats

• Cold outreach lists. You hold names and company domains and need the address shape before you build addresses.
• Account research. Check whether a target account uses first.last@ or flast@ before a campaign.
• List cleaning. Spot addresses that do not follow the company's own format.
• Agent workflows. An AI agent holds a domain and needs the address convention in a single call.

---

🚀 Quick start: email formats for a list of domains

1. One domain

{ "items": ["stripe.com"], "limit": 1 }

2. A list, in any form

{ "items": ["openai.com", "https://www.shopify.com/about", "sales@airbnb.com"], "sampleSize": 300 }

Each entry becomes its bare domain before the lookup.

---

⚙️ Inputs

Key	Type	Default	What it does
items	array of strings	required	Company domains, website URLs or email addresses. A company name is not a domain and is listed in MISSES as invalid_domain. The same domain written twice is looked up once. Free plan: 100 domains per run. Maximum 800 entries.
sampleSize	integer	200	How many of the company's current people are read for evidence, 50 to 500. Small companies are read in full whatever you set. The price does not change.
limit	integer	800	Most patterns delivered in the whole run, which also caps what is charged. Maximum 800. Domains not reached are listed in MISSES.
includeAudit	boolean	false	Adds an audit object to each row.

Any other field stops the run before a lookup with remove unsupported fields: .... Nothing is charged.

---

📤 Output

One result row per domain with a credible pattern. Rows keep input order. Empty values are left out.

Key	Type	Meaning, and when it is missing
domain	string	The domain from your entry, cleaned
pattern	string	The winning format, see the table below
emailFormat	string	The pattern written out: first@stripe.com
example	string	A masked address that follows the pattern
confirmations	integer	People whose address on record follows the pattern
agreement	number	confirmations divided by withAddress, 0 to 1
withAddress	integer	People read who have an address on the domain on record
sampled	integer	People read
coverage	number	withAddress divided by sampled, 0 to 1
otherPatterns	array	Other formats at least one person fits: {pattern, confirmations}, best first. Can be empty
samples	array	Up to three masked addresses on the domain
audit	object	Only with includeAudit: requestId, latencyMs, sampleSize, inputRaw

Example output (a real answer for stripe.com, 2026-11-06):

{
  "domain": "stripe.com",
  "pattern": "first",
  "emailFormat": "first@stripe.com",
  "example": "c*****@stripe.com",
  "confirmations": 50,
  "agreement": 0.595,
  "withAddress": 84,
  "sampled": 200,
  "coverage": 0.42,
  "otherPatterns": [
    { "pattern": "fl", "confirmations": 13 },
    { "pattern": "firstlast", "confirmations": 5 },
    { "pattern": "flast", "confirmations": 5 }
  ],
  "samples": ["c*****@stripe.com", "g*****@stripe.com", "b******@stripe.com"]
}

Pattern names

first is the first name, last the last name, f or l a first or last initial. So first.last is jane.doe@, flast is jdoe@, firstl is janed@, first_last is jane_doe@, last.first is doe.jane@. The data also names firstlast, first-last, f.last, first.l, lastfirst, last_first, lfirst, l.first, last.f, lastf, last, f_last and fl (taken as first initial plus last initial).

How to read the rows

• agreement is the number to watch. bbc.co.uk came back at 0.91 (first.last, 68 people), hubspot.com at 0.86 (flast), stripe.com at 0.60 (first). A company that changed its format over time or lets people pick an alias scores lower; the rows still return the winner with its evidence.
• It is a pattern read from addresses on record, not a delivery test. A pattern says how a company usually writes addresses. It does not prove that a given address exists or accepts mail. Verify before you send, with email-verifier.
• Large companies are read in part. The default reads 200 current people. coverage shows how many of them had an address on record, typically 30% to 70% in tests.
• Shared domains can mix companies. When a few companies declare one domain, the evidence comes from all of them. Domains declared by more than 50 companies (public mail, website builders) are refused for free (shared_domain).
• Small companies are often skipped. With fewer than 10 people with an address, the row is a free low_confidence, no_addresses or unresolved_domain notice.

Other records in the run

• MISSES (key-value store): [{index, input, reason}] for every entry without a result row. Reasons: the notice reasons above, invalid_domain, duplicate_input (the same domain as an earlier entry; duplicateOf names it), not_looked_up_limit (the run limit was reached first), or Upstream HTTP <code> / Upstream unavailable.
• errors-email-pattern-by-domain (key-value store): the lookups that failed for a technical reason, so you can re-run only those domains.
• RUN_SUMMARY: counts (requested, valid, looked up, patterns delivered, how many per pattern).
• Notice rows (type: "notice") appear for every free outcome, a free-plan list longer than 100 (free_plan_limit), a reached limit (limit_reached), a reached spend limit (charge_limit_reached) or invalid input (invalid-input). Notices are never charged.

What is NOT returned

• Real email addresses (samples are masked)
• A verified or deliverable address (use email-verifier)
• An address for a named person (use email-finder)
• The company behind a domain (use company-by-domain)

---

🛡️ Reliable results

Our data promise. We work to keep the data current, and we stand behind it. If a pattern looks wrong for a company you know, a company is missing or a row looks off, open the Issues tab with your run link. We review every report and sort it out with you, case by case.

• No data is free and visible. Every domain without a row is in MISSES with its position and a reason and is never charged.
• No guesses on thin evidence. Fewer than 5 confirming people, fewer than 10 addresses or less than 40% agreement returns a free notice.
• Repeats are charged once. The same domain written twice is one row.
• Bounded pace. Lookups run one after another, about 3 to 4 seconds each, so a long list takes time. The run timeout is one hour.
• Typos fail early. An unsupported input field, a bad sampleSize or limit or a list above 800 entries stops the run before any charge.

---

🤖 For AI agents (MCP)

This actor runs through the Apify MCP server. Prompt to copy:

"Use b2bdata/email-pattern-by-domain with items=["stripe.com"] and limit=1. Return pattern, emailFormat and agreement."

When to use it. You hold a company domain and need the shape of its work email addresses. When not to. You need a person's address (use b2bdata/email-finder), a deliverability check (use b2bdata/email-verifier) or the company (use b2bdata/company-by-domain). Input contract. items array of domains, URLs or emails (at most 800); optional sampleSize 50-500, optional limit 1-800. Output contract. One row per domain with domain, pattern, emailFormat, confirmations, agreement, withAddress, sampled. Output is deterministic and schema-stable. Domains without a credible pattern are a free notice row and listed in MISSES. Smoke test. limit: 1 on a well-known domain returns one row. Error cases. No valid domains means no entry was a domain. remove unsupported fields, a bad number or too many entries ends the run SUCCEEDED with one free invalid-input notice row and nothing charged.

---

🔌 Integrations: send results anywhere

1. API. Fetch rows from the run's default dataset: GET /v2/datasets/{datasetId}/items?format=csv.
2. Webhook on run success to push rows to your own endpoint.
3. n8n, Make, Zapier. Use the native Apify nodes: run the actor, then read the dataset.
4. Google Sheets, Airtable, HubSpot, Salesforce, Clay. Join on domain and use emailFormat to build addresses from names.

---

🕐 Schedule examples

• Weekly list prep. A schedule with a saved task that points at your target-account domain list.
• New-account enrichment. A webhook from your CRM starts a run with the new domains.

---

🔗 Related actors

• b2bdata/email-finder: the work email of one named person at a company. Open
• b2bdata/email-verifier: check whether an address is deliverable before you send. Open
• b2bdata/company-by-domain: the company behind a domain, with industry and size. Open
• b2bdata/people-at-company: the people at a company, one row per person, by domain. Open

See all actors by b2bdata

---

❓ FAQ

Data and coverage

How do I find the email format of a company? Put the company domain in items and run. You get one row with the pattern, such as first.last, and how many people confirm it.

Why did a domain return a notice instead of a pattern? No company is on file for it, nobody read has an address on it, the addresses follow no name-based format, or too few people confirm one format. All are free and are in MISSES with the reason.

Why is gmail.com refused? It is a public mail domain, declared by thousands of companies, so it has no company email format. It is a free shared_domain notice.

Is the pattern always right? No. It is the format most people on record follow. agreement shows how strongly. Companies with several formats, aliases or recent changes score lower. Verify addresses before you send.

Do I need a LinkedIn account or cookies? No. It is a lookup without login, cookies or an account to ban.

Pricing

What do I pay when a domain has no pattern? Nothing. There is no start fee, and unknown domains, skipped domains, thin patterns, bad entries, repeats and notices are free.

---

🛟 Support

Use the Issues tab on this actor's Apify page for bugs, questions and feature requests. We answer within 48 hours on working days. Please include the run ID and the input you used. Published on Apify by b2bdata. Not affiliated with, endorsed by, or sponsored by any company or person named in the examples above.