Markdown Link Extractor

Extract Markdown links and resolved reference destinations into a readable report without opening or fetching URLs.

Markdown is parsed in this browser. No extracted URLs are fetched. Input limit: 50,000 UTF-16 code units.

Includes parsed inline links, resolved references and GFM autolinks. Image destinations and raw HTML links are excluded.

Paste Markdown source and choose Inspect Markdown to list parsed link text, destination and one-based source line. The report includes inline links, resolved reference links and GFM autolinks. Destinations appear as text: the tool neither opens them nor checks whether a server responds.

Follow the built-in example

Choose Load example (replaces input), then inspect. Two rows appear: Guide points to https://example.com/guide on line 1, and Reference resolves to /help on line 2. The image destination is deliberately excluded. Reference definitions supply destinations but are not separate occurrences in the report.

Know which syntax is included

The existing remark parser with GFM support identifies links structurally, so link-like characters inside code are not reported. Inline formatting in a label is reduced to plain text. Reference identifiers follow the parser's normalization rules; if a definition is repeated, the first definition wins. An unresolved reference remains prose rather than a resolved link row.

Images, raw HTML anchors and raw HTML image tags are excluded. A linked image can still create a link row with its alternative text as the label. Relative destinations remain relative because no base website is supplied. Repeated occurrences remain separate, even if their URLs match. This is an inventory, not a broken-link checker, redirect resolver or security scanner. Treat unfamiliar destinations as untrusted text.

Review and export

The table previews the first 200 matching items. Copy JSON report and Download JSON include all matches in source order, with line, kind, text and detail fields. Zero links is a valid empty report. Input is limited to 50,000 UTF-16 code units, with 20,000 syntax nodes and 100 nesting levels. Browser textarea line endings may normalize before parsing. Edits clear the previous report.

Use Markdown Heading Extractor to inspect document structure, or URL Extractor when the input is ordinary text rather than Markdown syntax.