Scrape Emails From Website

Give it a company website, get back every email address published on it.

What it does

• Crawls a company website and extracts every published email address — page text, mailto: links, and embedded markup.
• Covers the homepage plus contact, about, team, support and legal pages.
• Returns generic inboxes — info@, sales@, support@ — far more often than named ones. Plan for that. A named address is a bonus, not the expected result.
• Crawling always starts at the site root. A deep link is reduced to its origin, so https://acme.com/pricing/enterprise is crawled as https://acme.com. You cannot point it at one page.
• Takes a website URL or a bare domain.

This is the fallback when there is no LinkedIn URL for find_work_email. Use it when you need any route into a company, not a specific person’s address.

Use cases

• Reach a company with no LinkedIn trail. Local businesses, agencies and SMBs often have a website and nothing else to enrich.
• Finish a Maps-sourced list. Listings give you sites; this turns those sites into addresses you can write to.
• Get any route into an account. A generic inbox that forwards internally beats no contact at all.
• Collect support and partnership addresses. Some motions are better served by partners@ or support@ than by a named rep.

How to use it

Scrape emails from acme.com.

Scrape emails from each of these 40 domains. Return one row per domain with
every address found. Do not guess addresses for domains that return nothing.

For anyone in this list where find_work_email returned nothing, scrape emails
from their company website instead and mark those rows as generic.

Verify every address you scraped and drop anything that isn't deliverable.

Example prompts

Best prompting practices

• Pass the bare domain. Deep links are reduced to the origin anyway, so acme.com saves the confusion. Lowercase, no www., no protocol.
• Expect generic inboxes. Tell your client to label the result as generic so it does not end up in a sequence written to a named person.
• Ask for the split. “Separate named addresses from role addresses — info@, sales@, bookings@ — and return both columns.” That split decides which rows are worth verifying and which get a different message.
• Forbid guessing. Add “do not guess email patterns” to any batch prompt. A fabricated first.last@company.com looks exactly like a scraped result and bounces like a fake.
• Reach for find_work_email first whenever you have a LinkedIn URL. It costs 1 credit and returns a named address instead of a shared inbox.
• Filter the list before the crawl. Drop rows with no website and domains you already own — free — before spending 0.5 credits a site. On 200 listings that is the difference between 100 credits and 40.
• Gate the batch. “Tell me how many sites survived the filter and what that costs, then wait for my go-ahead.”
• Verify separately. Published addresses go stale. Chain verify_email before you send.

Output

{
  "url": "https://acme.com",
  "emails": ["info@acme.com", "sales@acme.com"]
}

Credits and limits

0.5 credits per call. Run check_credits (free) before a large batch.

Parameters

Related tools

Next steps

• All MCP tools — the full catalogue with credit costs
• Research & Web Scraping — workflows that chain this tool into a list build

Keep learning

Two courses take these tools past the reference page — clustered by GTM job, then chained into workflows.