New: the Scrapier MCP server. Every scraper in Claude Code with one command. See how →
Free tool · no signup

Free URL Extractor

Get every link on a web page, sorted into internal, external, email and phone links, with the anchor text and nofollow flag. Copy the list or download a CSV.

One page per run · 30 runs/hour

By the Scrapier team · Free · No signup · Internal and external · CSV download · Last updated

Quick answerWhat is a URL extractor? A URL extractor reads a web page and lists every link on it as a full URL, removing duplicates. This one also labels each link as internal, external, email or phone, keeps the anchor text, and flags nofollow links, so you can filter and export them.
Sample

Sample result

The Wikipedia article "Web scraping" on 24 Oct 2026 returned 326 unique links in 95 ms: 224 internal and 102 external, 64 of them nofollow.

Input URLen.wikipedia.org/wiki/Web_scraping
ModeURL
Fetch time95 ms
Unique links326
Internal / external224 / 102
nofollow64
Skipped# anchors, javascript: links, duplicates
url,type,anchor_text,nofollow
"https://en.wikipedia.org/wiki/Main_Page","internal","Main page","false"
"https://en.wikipedia.org/wiki/Wikipedia:Contents","internal","Contents","false"
"https://en.wikipedia.org/wiki/Portal:Current_events","internal","Current events","false"
…
"https://donate.wikimedia.org/?wmf_source=donate&…","external","Donate","false"
"https://ca.wikipedia.org/wiki/Web_scraping","external","Català","false"
…
"https://en.wikipedia.org/wiki/Special:RecentChangesLinked/Web_scraping","internal","Related changes","true"
"https://en.wikipedia.org/wiki/Data_scraping","internal","Data scraping","false"

Excerpt of the CSV download. Language links to other Wikipedia editions count as external because their host differs. Run on .

How to

How to use the URL extractor

  1. Paste the page URL, or switch to Paste HTML to extract links from HTML.
  2. Press Extract URLs. Relative links are turned into full URLs and duplicates removed.
  3. Filter by type, then copy the URLs or download a CSV.
Output

What you get

Every link on the page as a full URL, one row each, with its type, anchor text and nofollow flag.

FieldWhat it containsExample
URLAbsolute URL, duplicates removedhttps://en.wikipedia.org/wiki/Data_scraping
Typeinternal, external, email (mailto:) or phone (tel:)internal
Anchor textThe clickable text of the link"Data scraping"
nofollowWhether the link has rel="nofollow"true on 64 of 326

Internal and external

Links to the same host are internal (www. is ignored); every other host is external. Relative links become full URLs.

Email and phone

mailto: links are listed as email and tel: links as phone. Anchors (#) and javascript: links are skipped.

Audience

Who it is for

SEO and content teams

Check internal linking, find outbound links, and spot nofollow links on a page.

Sales and growth teams

Collect the email and phone links a page publishes, or the outbound links a directory lists.

Developers

Get a page's links as CSV to seed a crawler, a link checker or a sitemap.

Uses

Common uses

  • Audit internal links on a page and spot pages that link out too much or too little.
  • Find outbound links and nofollow links before a content or link-building review.
  • Collect the email (mailto:) and phone (tel:) links a contact or directory page publishes.
  • Build a list of pages to crawl or convert, starting from a category page or an HTML sitemap.
  • Export a page's links to CSV for a spreadsheet, a link checker or a redirect map.
Alternatives

How it compares

This tool: Absolute URLs, typed, deduplicated, with anchor text. Watch out for: One page at a time. Other methods, and the case where each one fits:

View source and search for href

Good at: No tools. Watch out for: Slow; relative links and duplicates.

Browser extensions

Good at: One click. Watch out for: Install and permissions; export varies.

Site crawlers

Good at: Whole sites. Watch out for: Desktop apps or paid plans.

A Python script

Every link on any number of pages, written to CSV. You handle robots.txt, retries and parsing. The same job in a few lines of Python, for when you need to run it on many pages.

python
import csv, sys, requests
from urllib.parse import urljoin, urlparse
from bs4 import BeautifulSoup

url = "https://example.com"
soup = BeautifulSoup(requests.get(url, timeout=15).text, "html.parser")
host = urlparse(url).netloc
out = csv.writer(sys.stdout)
out.writerow(["url", "type", "text"])
for a in {a["href"]: a for a in soup.find_all("a", href=True)}.values():
    link = urljoin(url, a["href"])
    kind = "internal" if urlparse(link).netloc == host else "external"
    out.writerow([link, kind, a.get_text(strip=True)])

Scrapier API for platform data

Typed JSON from 35 platforms at volume, or inside an AI agent: see Use the API instead below.

MethodPagesAbsolute URLsTyped linksEffort
This tool1 per runYesYesNone
View source and search for href1NoNoManual search
Browser extensions1VariesVariesInstall
Site crawlersWhole sitesYesYesDesktop app or paid plan
Limits

Limits and tips

  • Filter by type before copying to get only internal URLs, only external URLs, or only email links.
  • To list pages across a whole site, run the extractor on the HTML sitemap or the category pages, then on the pages it finds.
  • Links added by JavaScript after the page loads will not appear; paste the rendered HTML instead.
  • Reads the HTML a server sends. Pages that build their content with JavaScript in the browser can come back short or empty; paste the rendered HTML instead (in your browser: right-click, Inspect, copy the <html> element).
  • Public pages only: no logins, paywalls or private content.
  • Respects robots.txt. If a site asks automated tools not to fetch a page, the tool stops and says so.
  • One page per run, up to 3 MB of HTML. 30 runs per hour per visitor.
API

Use the API instead

The free tool reads one page at a time. Scrapier's API returns typed JSON from Instagram, TikTok, Amazon, Google Maps and 31 more platforms: 199 endpoints, billed per successful call. It does not fetch arbitrary pages; it gives you the platform data directly, over REST or as an MCP tool inside an AI agent.

/google-search/
curl -G "https://proxy.scrapier.io/api/v2/google-search/" \
  -H "X-API-Key: YOUR_API_KEY" \
  --data-urlencode "queries=web scraping api"
# → typed JSON, 1 credit per successful call; failed calls are free
MCP · example run

"Find the top 10 Google results for ‘web scraping api’ and list their URLs."

search_endpoints → run_scrape /google-search/ × 1

→ 1 successful call, 1 credit each: 1 credit. Failed calls are free.

Free toolScrapier API
ReadsAny public web page, or HTML you paste199 endpoints on 35 platforms
OutputTyped link list, CSVTyped JSON per endpoint
Pages1 per runMany, up to 3 parallel requests per key
Rate limit30 runs/hour per visitor, 3 MB per pageYour plan's credits; 429 with Retry-After if you go too fast
SignupNoneAPI key; a free tier is available
PriceFree, no signupPer successful call, failed calls free. From $29/mo for 15,000 credits

Safe and legal use

Fetching a public page is the same request a browser makes. The tool respects robots.txt, refuses private and local network addresses, and does not store the pages it reads or the HTML you paste. The site's terms and copyright govern how you reuse what you extract. Not legal advice; see our Terms of Use and Privacy Policy.

FAQ

Frequently asked questions

Is the URL extractor free?

Yes. No signup and no charge, up to 30 runs an hour per visitor.

What links does it extract?

Every <a href> on the page: internal and external links, mailto: email links and tel: phone links. Anchors (#) and javascript: links are skipped.

Can it find broken links?

It lists the links but does not test them. Download the CSV and check the status codes with a link checker.

Can I extract links from a whole website?

The tool reads one page per run. Start from the homepage or the sitemap page, then run it on the pages you need.

Can I export the results?

Yes. Copy the URLs to the clipboard, or download a CSV with URL, type, anchor text and nofollow.

Does it work on JavaScript-heavy pages?

It reads the HTML the server sends. Links that a page adds with JavaScript after loading will not appear; paste the rendered HTML instead.

Scale

Structured data from 35 platforms, one key.

199 endpoints that return typed JSON. Billed per successful call; failed calls are free.