URL Extractor: Extract URLs, Domains & Emails From Text

What it can extract

Option	From `Read https://blog.example.co.uk/post?utm_source=x or mail jo@shop.example.com`	Use it for
URLs	`https://blog.example.co.uk/post?utm_source=x`	Link lists, crawling, redirect maps, checking pages in Search Console
Domains	`blog.example.co.uk`, `shop.example.com`	Seeing every host that’s mentioned, including subdomains
Root domains	`example.co.uk`, `example.com`	Counting sites, disavow files, outreach lists, company names from work emails
Emails	`jo@shop.example.com`	Contact lists from exports and old pages

You can tick several at once. Each gets its own section in the result, or its own column in CSV and its own key in JSON.

5 jobs I use it for

1. Find every external link on a page

Open the page and press Ctrl+U (Cmd+Option+U on a Mac) to see its source.
Select all, copy, and paste it into the box.
Under Options, type your domain in Your site and choose External only.
Tick Skip images, CSS, JS and font files to keep only real links.

This is how I check a post before publishing: no forgotten staging links, no links to a competitor, and nothing pointing to a dead site. With Your site filled in, relative links like /about/ are turned into full URLs too.

2. Turn a backlink export into a disavow file

Paste the CSV from Search Console (Links → Export external links), Ahrefs, Semrush or any other tool. Tick Root domains, set Your site with External only so your own pages are left out, and choose Google disavow file as the format. You get one domain:example.com line per site, ready to upload to Google’s disavow tool. If you also tick URLs, links on sites that are already disavowed as a whole are left out.

Be careful with it. Google says most sites don’t need a disavow file, because it already ignores most spammy links. Use one only for links you built or paid for, or links named in a manual action, and read the list line by line before you upload it. The sample “Backlink export” above shows why: bbc.co.uk is in it, and you would never want to disavow that.

3. Get all URLs from a sitemap

Open yoursite.com/sitemap.xml (or sitemap_index.xml with Yoast), view the source, and paste it. Every <loc> becomes one line. For Persian, Arabic or other non-English slugs, tick Make %-encoded URLs readable so %D8%B3%D9%84%D8%A7%D9%85 shows as the actual word. Then use Only keep results containing with /blog/ or /product/ to pull out one section.

4. Clean a list of links before sharing it

Links copied from newsletters, ads and Google Shopping carry tags like utm_source, gclid, fbclid and srsltid. Remove tracking parameters strips those and keeps the parameters the page needs, such as ?id=5. Unwrap redirect links turns google.com/url?q=…, Facebook’s l.php?u=… and Outlook “safe links” back into the real address.

5. See which sites a text links to most

Under the result, Top sites in this text counts links per site. Click a site to see only its links. I use this on Telegram chat exports (Telegram Desktop → Export chat history → JSON) to see which websites a group shares most, and on long guest-post pitches to spot the one site every “article” links to.

How it decides what counts as a link

Most extractors either miss links without http:// or flag every file.txt as a website. This one follows a few rules, so you can trust the list:

Real endings only. A name like example.com is a link only if .com is in IANA’s official list of top-level domains (1,436 of them as of November 2026). That rules out index.php, image.png, node.js and v1.2.3.
File names that look like domains. .md, .py, .sh, .ai, .zip and .mov are real country or brand endings but also common file extensions. Without http:// or www., they only count when a path follows, so readme.md is skipped but nic.ai/whois is found.
Sentence punctuation is dropped. A full stop, comma, quote or the Persian ، after a link isn’t part of it. A closing bracket is kept only if the link opened one, so (see https://en.wikipedia.org/wiki/Python_(programming_language)) comes out right.
Encoded links are decoded first. & in HTML and https:\/\/ in JSON both become normal links.
Emails come before domains. info@site.com is read as an email, not as a link to site.com. If you want the domains of email addresses, keep Include domains from email addresses on.

Domain vs root domain

The root domain is the part someone actually registers. You can’t simply take the last two parts, because in the UK, Australia, Iran and many other countries people register under a second level such as .co.uk, .com.au or .co.ir. The tool uses the multi-part endings from the Public Suffix List, the same list browsers use to decide which sites can share cookies.

Host	Root domain
`news.bbc.co.uk`	`bbc.co.uk`
`shop.example.com.au`	`example.com.au`
`www.example.co.ir`	`example.co.ir`
`mysite.blogspot.com`	`mysite.blogspot.com` (each blog is its own site)
`192.168.1.10`	`192.168.1.10`

Blogs on blogspot.com, wordpress.com, github.io, netlify.app and similar platforms belong to different people, so by default each subdomain counts as its own site. That matters for disavow files: you want to disavow spam.blogspot.com, not all of Blogger. Untick Treat blog platforms as separate sites to group them.

All options

Option	What it does
Find links without http://	Finds `example.com` and `www.example.com/page` in plain text. Add https:// gives them a full address in the result.
Read links in HTML attributes	Reads `href` and `src`, including `//cdn.example.com` links and, when Your site is set, relative links.
Your site + All / External / Internal	Internal means the same root domain, so `blog.yoursite.com` counts as internal.
Only keep / Remove results containing	Comma-separated words, matched anywhere in the result, ignoring case.
Remove duplicates	On by default. Show how many times each appears adds the count.
Sort	Order found, A to Z (ignoring `https://` and `www.`), most frequent first, or grouped by site.
Show international domains as xn--	Shows `xn--bcher-kva.de` instead of `bücher.de`, for tools that only accept ASCII.
Find hidden emails	Reads `name [at] site [dot] com` and `name (at) site (dot) com`.
Refang defanged links	For security reports: `hxxps://evil[.]com` becomes `https://evil.com`. Don’t open those links.

Tested November 2026 against 56 test cases, including Wikipedia links with brackets, Persian slugs and punctuation, JSON exports, HTML entities and file names that look like domains. A 5.3 MB file with 100,000 links was processed in about one second. Above one million characters the result no longer updates as you type; press Extract instead.

FAQ

Is what I paste uploaded anywhere?

No. Everything runs in your browser; nothing you paste or open is sent to DanialTV. You can disconnect from the internet after the page loads and it still works. Only your option choices are remembered in your browser, not the text.

Can it fetch a web page and extract its links?

Not directly. Browsers don’t let one site read another site’s pages, and doing it through a server would mean your data passes through mine. Open the page, press Ctrl+U, copy the source and paste it here. It takes five seconds.

Does it work with PDF, Word or Excel files?

The Open file button reads text files: TXT, CSV, TSV, HTML, XML, JSON, Markdown and logs, up to 50 MB each and 20 at a time. For PDF, Word or Excel, select all the content, copy it and paste it in. For Excel, saving the sheet as CSV first works best.

Why are some domains missing?

Check three things: the ending must be a real top-level domain, a file-like ending such as .md needs a path after it, and Find links without http:// must be on. Names on internal networks, such as http://intranet/, are found only when they start with http:// or https://.

More free tools: the robots.txt tester checks whether Google can crawl the URLs you extracted, and the Search Console regex builder turns a list of pages into a filter for your performance report.