Link Extractor
Paste HTML or Markdown and pull out every hyperlink and in-page anchor — with link text and a relative-link flag. Runs entirely in your browser, no upload, no sign-up.
Extract links from HTML or Markdown in your browser
Paste HTML or Markdown and get back every hyperlink and every in-page anchor (jump target). Each link comes with its destination URL, its visible link text, and a flag marking relative links. Everything runs locally in your browser — your input is never uploaded to a server.
What it finds
- Hyperlinks — HTML
<a href>and<area href>; Markdown inline links, reference links, and autolinks (<https://…>). For each one you get the URL and the link text. - Anchors (jump targets) — HTML
idand legacy<a name>attributes; Markdown heading slugs (GitHub-style, e.g.## Big Section→big-section) and any explicit<a name="…">/id="…"anchors embedded in the Markdown. These are the targets a same-document#namelink jumps to.
Options
- Parse as —
auto(default) sniffs HTML vs. Markdown from the content; force it withhtmlormarkdownif the auto-detection guesses wrong. - Base URL — optional. Give an absolute URL like
https://example.com/docs/and every relative link is resolved to its full absolute form (absolute, fragment, andmailto:links are left untouched). - Remove duplicate links — collapse links that point at the same final URL, keeping the first occurrence.
- Output —
text(default) for a readable list, orjsonfor a structured{ source, link_count, anchor_count, links, anchors }object you can pipe into other tools. Each JSON link carries itsurl,text, arelativeflag, and any HTMLrelattribute (e.g.nofollow,sponsored).
Notes
- Link text is whitespace-collapsed (newlines and runs of spaces become a single space).
- Relative links (
/path,./a,docs/x.md) are flagged; fragment-only (#section), protocol-relative (//cdn…), and scheme-bearing (https:,mailto:) destinations are reported as written and counted as absolute. - URLs are reported verbatim — they are not resolved against a base URL.
FAQ
The auto-detection parsed my input as the wrong format — what do I do?
Set Parse as to html or markdown explicitly instead of leaving it on
auto. Mixed content (Markdown with embedded HTML tags) is the usual trigger for
a wrong guess. If you're not sure which parser ran, the JSON output's source
field tells you which one auto-detection actually picked.
How do I turn relative links like /docs/page into full URLs?
Fill in the Base URL field with an absolute URL such as
https://example.com/docs/. Every relative link is then resolved against it and
its relative flag is cleared. Links that are already absolute — including
fragment-only (#section), protocol-relative (//cdn…), and mailto:
destinations — are left exactly as written.
What counts as an "anchor" in the results?
Anything a same-document #name link can jump to: HTML id="…" attributes,
legacy <a name="…"> anchors, and — in Markdown — GitHub-style heading slugs, so
## Big Section is reported as the anchor big-section. Anchors are listed
separately from hyperlinks with their own count.
Does "remove duplicate links" merge links with different text?
It collapses by destination URL: when several links point at the same final
URL, only the first occurrence is kept, whatever its text. It's off by default,
so an unfiltered run shows every <a> tag — handy when you're auditing anchor
text rather than collecting URLs.
Developer & Automation Access
Run it from the terminal
Same engine as this page, headless — via the gizza CLI:
gizza tool link-extractor "Paste HTML or Markdown containing links…"New to the CLI? Get gizza →
Open it by URL
Pre-fill and auto-run this tool with query parameters — the names match the API/CLI:
https://gizza.ai/tools/link-extractor/?input=Paste%20HTML%20or%20Markdown%20containing%20links%E2%80%A6&source=auto&base_url=https%3A%2F%2Fexample.com%2Fdocs%2F&dedup=true&output=textMachine-readable descriptor: tool.json — title + parameters JSON Schema for agents.