When a business website shows [email protected], or its email link points to /cdn-cgi/l/email-protection#..., the site is using Cloudflare's Email Address Obfuscation. Cloudflare replaces each address in the HTML with an encoded link and adds a small script that restores the real address in a visitor's browser. People see the email normally; tools that only read the raw HTML see the placeholder, which is why many email scrapers report "no email found" on sites that clearly show one.
This explainer covers what visitors and scrapers each see, how the encoding works at a high level, the other ways sites hide addresses, and what to do when the contact form is the only route. It belongs to our email finding cluster; the pillar is how to find email addresses from a domain.
What visitors see and what scrapers see
Cloudflare's documentation says the feature hides addresses "from bots, while keeping them visible to humans", that "there are no visible changes to your website for visitors", and that it is enabled automatically when a site signs up.
For a fictional contact page on oakline-joinery.example:
| Who is reading | What they get |
|---|---|
| A person in a normal browser | "Email us at [email protected]", clickable |
| A person with JavaScript turned off | "Email us at [email protected]" |
| A tool that downloads the HTML and searches it | An <a> tag linking to /cdn-cgi/l/email-protection# followed by a long hex string, with the text [email protected] |
| A crawler that renders the page in a browser | The same as the person: the real address |
| Someone copying text from search results | Often the placeholder, because search snippets come from the HTML |
That last row explains why people paste [email protected] into search engines: they have copied it from a snippet or a scraped list and are trying to find out what it means.
How the encoding works, at a high level
Three pieces work together:
- Replacement. When Cloudflare serves the page, it finds email addresses in the HTML and swaps each for a link to
/cdn-cgi/l/email-protectionplus an encoded version of the address. The visible text becomes[email protected]. - The encoded string. The address is stored as a hex string. An independent analysis published in 2018 showed that the first byte acts as a key and each following byte is combined with that key to recover one character. The author describes this as cryptographically weak, and notes Cloudflare's own position that it is meant to obfuscate addresses, not keep them confidential.
- The decode script. Cloudflare injects a small deferred script (
email-decode.min.js) that runs as the page loads, finds the encoded links and writes the real addresses back into the page.
The point for anyone collecting business contacts is not the maths. It is that the address is still published to every human visitor; it is simply not in plain text in the HTML. Anything that behaves like a visitor, by loading the page and running its scripts, sees it.
Be honest about what the feature means, though. Cloudflare presents it as a way to hide addresses from bots, so it is a signal about automated harvesting, even though many owners never chose it (it is on by default). Read the address the way a visitor would, at a visitor's pace, from pages a visitor would open, and do not build tools whose only purpose is to strip the protection from pages at scale. What you then do with the address is governed by the same marketing and data protection rules as any other business contact.
Where Cloudflare does not obfuscate
Cloudflare lists cases the feature skips, which explains why some addresses on a Cloudflare site are plain and others are hidden:
- Addresses in HTML attributes, except the
hrefof a link. - Addresses inside
<script>,<noscript>,<textarea>,<xmp>or<head>tags. - Pages that are not served as HTML (for example JSON responses or PDFs).
- Responses with a
Cache-Control: no-transformheader. - Content added by a Cloudflare Worker.
- Addresses wrapped in
<!--email_off-->and<!--/email_off-->comments.
So structured data in the <head>, such as a Schema.org email property inside JSON-LD, is often left in plain text even when the visible address is obfuscated.
Checking a page yourself
You do not need any tools to confirm what is going on with a single site:
- Open the contact page normally. If an address is visible, it is published.
- View the page source (
Ctrl+U, orCmd+Option+Uon a Mac) and search foremail-protection. If you find it, the site uses Cloudflare obfuscation for that address. - Search the source for
ld+json. If the business's structured data includes anemailproperty, it is often in plain text there. - If the address appears on screen but neither the source nor the structured data contains it, a script is building it after the page loads.
Whichever case applies, the address on screen is the one the business chose to show visitors, and it is the one to record, with the page URL as its source.
Other obfuscation styles
Cloudflare is the most common case, but not the only one.
| Style | Example | Readable by a person | Readable by an HTML-only extractor | Readable by a rendered crawl |
|---|---|---|---|---|
| Cloudflare obfuscation | [email protected] | Yes | No | Yes |
| Words for symbols | hello [at] oakline-joinery [dot] example | Yes | Only if it understands the convention | Only if it understands the convention |
| Reversed or split text with CSS | Characters reversed and flipped back with styling | Yes | No | Sometimes |
| JavaScript assembly | A script joins "hello", "@" and the domain | Yes | No | Yes |
| HTML entities | hello@... | Yes | Yes, after decoding entities | Yes |
| Image of the address | A PNG showing the email | Yes | No | No |
| Contact form only | No address published | No address to read | No | No |
The words-for-symbols style is common on small business sites where the owner typed it by hand. It is easy to read, but extractors need explicit rules for the variations: (at), [at], at, [dot], dot, and local language equivalents.
An image of an address is a clear signal that the owner does not want automated collection. Do not try to read it automatically; if you need the address, note it by hand from a page you are visiting anyway, or use the form.
When the only route is the contact form
Some businesses publish no address at all. The contact page has a form, perhaps a phone number, and nothing else. A deeper crawl will not change that.
Treat this as information, not failure:
- Record it. A "forms only" status in your data tells you the channel to use.
- Use the form as intended, with a short, relevant, honest message from an identified sender. Many businesses read form submissions more closely than a general inbox.
- Do not hunt for a hidden address in WHOIS, old directory listings or guessed patterns. The business has made a choice about how it wants to be contacted.
Contact form outreach has its own etiquette and the same legal basics as email. In the UK, the ICO's B2B guidance says you must identify yourself and offer a way to opt out even when PECR's consent rule does not apply to corporate subscribers; sole traders and some partnerships are individuals who need consent or the soft opt-in for email marketing. Other countries differ.
A note on fairness
Cloudflare's feature exists to reduce automated harvesting for spam. Recovering an address that a site shows to every visitor, by rendering the page as a browser does, is different from defeating an access control, and it is the same thing a person does when they read the page. Still, the reason the owner turned the feature on (or left it on by default) is worth respecting: collect only the business's published contact details, only for a purpose you can justify, respect the site's terms and robots.txt, and keep your request rate low. Do not use these techniques to bypass CAPTCHAs, logins, rate limits or other measures that actually restrict access.
How Map Extract helps
The email extractor is designed to decode common address obfuscation on the pages it crawls, and the website data extractor underneath it has a rendered browser fallback for sites whose content depends on scripts, so addresses that appear on screen after the page loads can be captured. It reads structured data as well as visible text and links, which catches JSON-LD email properties that Cloudflare leaves in plain text.
Each address is stored with the page it was found on and an inbox purpose label, so you can see whether it was the contact page address or a privacy contact. Sites that block the crawler or show a CAPTCHA are recorded and quarantined, not worked around.
What it does not do: addresses published only as images are likely to be missed, it cannot find an address on a forms-only site, and decoding coverage depends on the obfuscation a site uses, so check the source column if a known address is missing. It also does not decide whether you may email what it finds.
Public business and social results the app collects are uploaded automatically to Map Extract's servers while it is licensed and running; the privacy notice explains this.
Related guides
- How to find email addresses from a domain
- Extract emails from a list of websites
- Extract email addresses from text
- Email list cleaner
Key takeaways
[email protected]and/cdn-cgi/l/email-protectionmean Cloudflare has replaced the address in the HTML.- Visitors still see the address because a small script restores it in the browser.
- HTML-only extractors miss it; rendered crawls and extractors that understand the encoding do not.
- Structured data in the page head is often left unobfuscated.
- Images of addresses and forms-only sites are deliberate choices; use the form rather than hunting.
Frequently asked questions
What does [email protected] mean on a website?
It is a placeholder inserted by Cloudflare's Email Address Obfuscation feature. Cloudflare replaces each email address in the page's HTML with an encoded link and a small script that turns it back into the real address in the visitor's browser. If you see the placeholder, the page's script did not run, for example because JavaScript is off or you are reading the raw HTML.
What is /cdn-cgi/l/email-protection?
It is the path Cloudflare uses for links that replace protected email addresses. The real address is stored in an encoded string attached to the link, and Cloudflare's decode script restores it when the page loads in a browser. Clicking the link with scripts enabled opens your mail client as a normal mailto link would.
Why does my email scraper miss addresses on Cloudflare sites?
Most simple extractors download the HTML and search it for text shaped like an email address. On a site using Cloudflare obfuscation, that text has been replaced with a placeholder and an encoded string, so there is nothing to match. A crawler that renders the page in a real browser, or that understands the encoding, sees the same address a visitor sees.
How do I stop Cloudflare hiding my own email address?
Cloudflare enables Email Address Obfuscation automatically for new sites. You can switch it off for the whole zone in the dashboard's security settings, disable it for specific hostnames or paths with a configuration rule, or wrap a single address in email_off comment tags so Cloudflare leaves it alone.
Sources
Written by the Map Extract editorial team. Found an error? Email [email protected] and we will check and correct it. How we research guides: editorial policy.