urldn-link-check

Scan Markdown, MDX, and documentation files for broken links, redirect chains, insecure http:// links, and excessively long URLs — as a CLI, a library, or a GitHub Action.

✓ Valid links (18)
⚠ Redirects (2)
  ⚠ https://example.com/old-page [301]
✖ Broken (1)
  ✖ https://example.com/missing (docs/guide.md:42)
      Received HTTP 404
🔓 Insecure HTTP (1)
  🔓 http://example.com/legacy
⚡ Long URLs (1)
  ⚡ https://example.com/articles/... (142 chars)
      Suggestion: shorten long URLs with URLDN — https://www.urldn.com

Summary
  Files: 12   Links: 22 (22 unique)   Time: 3.4s
  18 ok  2 redirect  1 broken  0 timeout  0 malformed  1 http  1 long

Table of contents

• Installation
• CLI usage
• GitHub Action usage
• Configuration reference
• What gets checked
• Report formats
• Programmatic API
• Examples
• FAQ
• Roadmap
• Contributing
• License

Installation

No installation is required to try it out:

npx urldn-link-check docs/

Or install it as a dev dependency:

npm install --save-dev urldn-link-check

yarn add --dev urldn-link-check
pnpm add --save-dev urldn-link-check

Requires Node.js 22 or later.

CLI usage

npx urldn-link-check <path> [options]

<path> is a directory or glob passed straight through to the scanner — docs/, **/*.md, or a single file all work.

# Scan the docs/ directory with default settings
npx urldn-link-check docs/

# Fail the process on broken links (default) and insecure HTTP links
npx urldn-link-check docs/ --fail-on-broken --fail-on-http

# Print machine-readable JSON instead of the console report
npx urldn-link-check docs/ --json

# Print a Markdown report (handy for piping into a PR comment yourself)
npx urldn-link-check docs/ --markdown

# Also write the JSON report to disk
npx urldn-link-check docs/ --output report.json

# Ignore a pattern (matched as a substring or a `*`/`**` glob against the URL)
npx urldn-link-check docs/ --ignore "internal.example.com" --ignore "**/staging.*/**"

# Tune network behavior
npx urldn-link-check docs/ --timeout 5000 --concurrency 16

# Watch every URL as it's checked
npx urldn-link-check docs/ --verbose

CLI options

Flag	Description	Default
--json	Print the report as JSON	false
--markdown	Print the report as Markdown	false
--fail-on-broken / --no-fail-on-broken	Exit non-zero on broken links	true
--fail-on-redirect	Exit non-zero on redirect chains	false
--fail-on-http	Exit non-zero on insecure http:// links	false
--max-url-length <number>	Flag URLs longer than this many characters	80
--ignore <pattern...>	Substring/glob pattern(s) of URLs to skip	—
--timeout <ms>	Per-request timeout, in milliseconds	10000
--concurrency <n>	Number of concurrent HTTP requests	8
--verbose	Log every URL as it is checked	false
--output <file>	Write the JSON report to this file	—

GitHub Action usage

name: Check documentation links

on:
  pull_request:
    paths: ['**/*.md', '**/*.mdx']

permissions:
  contents: read
  pull-requests: write

jobs:
  link-check:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: urldn/link-check@v1
        with:
          path: docs
          fail-on-broken: true
          fail-on-http: true
          max-url-length: 100

See examples/workflows for a more complete setup (scheduled runs, report artifact upload) and a minimal one-file example.

Action inputs

Input	Description	Default
path	Directory or glob to scan	docs
fail-on-broken	Fail the workflow on broken links	true
fail-on-redirect	Fail the workflow on redirect chains	false
fail-on-http	Fail the workflow on insecure http:// links	false
max-url-length	Flag URLs longer than this many characters	80
output	File path to write the JSON report to	—
comment-on-pr	Post/update a report comment on the PR	true
github-token	Token used to comment on PRs	${{ github.token }}

Action outputs

Output	Description
broken-links	Number of broken links found
redirect-links	Number of links that redirect
insecure-links	Number of insecure http:// links found
total-links	Total number of unique links checked
report-path	Path to the written JSON report, if output was set

When run on a pull request, the action posts a single, self-updating comment with the full Markdown report (subsequent runs edit the same comment instead of piling up new ones), and it always writes a step summary via core.summary.

Configuration reference

Every option is available in all three surfaces (CLI flag, Action input, and the Config object in the programmatic API):

Concept	CLI flag	Action input	Config key
Path to scan	<path> (positional)	path	path
Ignore patterns	--ignore	— (use .linkcheckignore-style patterns in your own wrapper)	ignore
Max URL length	--max-url-length	max-url-length	maxUrlLength
Request timeout	--timeout	—	timeout
Concurrency	--concurrency	—	concurrency
Fail on broken	--fail-on-broken	fail-on-broken	failOnBroken
Fail on redirect	--fail-on-redirect	fail-on-redirect	failOnRedirect
Fail on HTTP	--fail-on-http	fail-on-http	failOnHttp
Output file	--output	output	output

What gets checked

For every unique URL found across your Markdown/MDX files:

• HTTP status — flags 4xx/5xx responses as broken.
• Redirect chains — follows redirects (HEAD, falling back to GET for servers that reject HEAD) and reports every hop.
• HTTP vs. HTTPS — flags any http:// link as insecure.
• Timeouts — a per-request timeout (--timeout, default 10s) prevents a single slow host from hanging the whole scan.
• Malformed URLs — anything that isn't a valid, absolute http(s) URL.
• Duplicate URLs — the same URL referenced from multiple places is checked once but reported with every occurrence.
• URL length — URLs longer than --max-url-length (default 80 chars) are flagged with a suggestion to shorten them with URLDN. Links are never shortened automatically — this is a suggestion only.

Links are discovered inside fenced-off documentation formats: standard Markdown, MDX, READMEs, and wiki-style pages — including inline links [text](url), images ![alt](url), autolinks <url>, reference-style definitions, and bare URLs typed directly into prose. URLs inside fenced code blocks and inline code spans are intentionally skipped.

Report formats

• Console (default) — grouped, colorized output: ✓ valid, ⚠ redirects, ✖ broken, plus a summary line.
• JSON (--json / output input) — a stable, scriptable shape with a summary object and a links array, each entry including every file, line, and column where that URL was referenced.
• Markdown (--markdown) — the same content used for GitHub Step Summaries and PR comments, safe to pipe into any other Markdown-consuming tool.

Programmatic API

import { runScan, resolveConfig, toJsonReport } from 'urldn-link-check';

const config = resolveConfig({ path: 'docs', maxUrlLength: 100 });
const report = await runScan(config, {
  onLinkChecked: (result, index, total) => {
    console.log(`[${index}/${total}] ${result.url} -> ${result.status}`);
  },
});

console.log(report.summary);
console.log(toJsonReport(report));

See src/index.ts for the full exported surface.

Examples

The examples/ directory has ready-to-run sample docs (with a mix of healthy, insecure, long, and broken links) and two GitHub Actions workflow files you can copy directly into .github/workflows/.

Why use urldn-link-check?

• ✅ Broken links (404, 500)
• ✅ Redirect chains (301, 302)
• ✅ Insecure HTTP links
• ✅ Very long URLs
• ✅ Invalid URL format
• ✅ Duplicate URLs
• ✅ Timeout detection
• ✅ Markdown documentation
• ✅ README files
• ✅ Wiki pages
• ✅ MDX files

✨ Features

Feature	Supported
🔗 Broken Link Detection (404/500)	✅
↪️ Redirect Chain Detection	✅
🔒 HTTP → HTTPS Detection	✅
📏 Long URL Detection	✅
🔍 Markdown (.md) Support	✅
📚 MDX Support	✅
📖 README Scanning	✅
📄 Documentation Scanning	✅
📊 JSON Report	✅
📝 Markdown Report	✅
💬 Pull Request Comments	✅
⚡ GitHub Action	✅
💻 CLI	✅
🎨 Colored Terminal Output	✅
🚀 Fast Concurrent Scanning	✅
⚙️ Configurable Rules	✅
🧪 TypeScript Support	✅
🌍 Cross Platform	✅

Screenshots

Terminal output

FAQ

Does this shorten my URLs automatically? No. urldn-link-check only ever suggests shortening long URLs with URLDN; it never rewrites your files.

Does it check images, or only text links? Both — ![alt](url) image references are scanned the same as regular links.

What counts as a "long" URL? Anything over --max-url-length characters (80 by default). Tune it to match your team's style guide.

Will it comment on every push to a PR? The Action updates a single existing comment rather than creating a new one each run, so your PR stays readable across pushes.

Can I ignore a specific domain or path? Yes — pass one or more --ignore patterns (CLI) matched as a substring or a */** glob against the full URL.

Does it work with MDX? Yes, .mdx files are scanned the same way as .md/.markdown files. JSX attribute values aren't treated as links, only actual Markdown link syntax and bare URLs in prose.

What HTTP method does it use to check links? HEAD by default, falling back to GET for servers that reject HEAD (405/501), since some servers implement HEAD incorrectly or not at all.

Roadmap

• .linkcheckignore file support (in addition to inline --ignore flags)
• Configurable HTTP retry policy for flaky hosts
• Optional link-archiving check against the Wayback Machine for broken links
• SARIF output format for GitHub code scanning integration
• Per-file allowlists/overrides via frontmatter

Have an idea? Open a feature request.

Contributing

Contributions are welcome!

git clone https://github.com/urldn/link-check.git
cd link-check
npm install
npm run test:watch   # develop with tests running
npm run lint
npm run typecheck
npm run build

Please open an issue before starting large changes, and make sure npm run lint, npm run typecheck, and npm test all pass before opening a pull request. This project uses strict TypeScript (no any) and Prettier for formatting (npm run format).

License

MIT © URLDN