Check if urls are dead or alive.

• What is this?
• When should I use this?
• Install
• Use
• API
  • deadOrAlive(href[, options])
  • defaultAnchorAllowlist
  • defaultSleep(retries)
  • AnchorAllow
  • Options
  • Result
  • Sleep
• Errors
  • dead
  • fetch
  • lost-hash-with-meta-http-equiv
  • lost-hash-with-non-html
  • lost-hash-with-redirect
  • max-redirect
  • missing-anchor
  • shared-declarative-refresh
• Compatibility
• Security
• Contribute
• License

The main goal of this project is to make sure URLs in docs are okay. As in, “if I link to it and a user goes there, will it work, and will they be there or somewhere else?” It uses modern JavaScript that works everywhere, and it uses the DOM when available, so you can use it in browsers, Node.js, workers.

It is made to be used by other tools. Such as a markdown tool, remark-lint-no-dead-urls.

• works in browser, node, worker, etc
• checks if anchors (such as #readme) point to elements
• follows HTTP and HTML redirects
• supports DOM clobber prefix (user-content-)
• emits warnings

You can use this when you want to build a link checker.

If you want to support SPAs or other client-side JS, this doesn’t do that.

This package is ESM only. In Node.js (version 18+), install with npm:

npm install dead-or-alive

In Deno with esm.sh:

import {deadOrAlive} from 'https://esm.sh/dead-or-alive@1'

In browsers with esm.sh:

<script type="module">
  import {deadOrAlive} from 'https://esm.sh/dead-or-alive@1?bundle'
</script>

import {deadOrAlive} from 'dead-or-alive'

console.log(await deadOrAlive('https://something-that-is-dead.com'))
console.log(await deadOrAlive('https://github.com'))

…yields:

{
  messages: [
    [1:1: Unexpected error fetching `https://something-that-is-dead.com/`] { … }
  ],
  status: 'dead',
  url: undefined,
  urls: undefined
}
{
  messages: [],
  status: 'alive',
  url: 'https://github.com/',
  urls: Set(203) {
    'https://github.githubassets.com/',
    'https://avatars.githubusercontent.com/',
    …
  }
}

This package exports the identifiers deadOrAlive, defaultAnchorAllowlist, and defaultSleep. It exports the TypeScript types AnchorAllow, Options, Result, and Sleep. There is no default export.

Check if a url is dead or alive.

• href (URL or string) — URL
• options (Options, optional) — configuration

Result (Promise<Result>).

To improve performance, decrease maxRetries and/or decrease the value used for sleep. The normal behavior is to assume connections might be flakey and to sleep a while and retry a couple times.

If you do not care about HTML redirects, whether anchors work, and what further URLs are used on, you can pass checkAnchor: false, findUrls: false, and followMetaHttpEquiv: false, which enables a fast path without parsing HTML.

Allow certain anchors (Array<AnchorAllow>).

This currently allows text fragments everywhere.

Calculate milliseconds to sleep between tries.

The function is defined as x ** 3 * 1000, so the first sleep is 1 ** 3 * 1000 is 1s, 2nd is 8s, 3rd is 27s, etc.

• retries (number) — try

Milliseconds to sleep (number).

Allow extra anchors (TypeScript type).

The first item is a regular expression to match URLs (origin and path, so without search or hash), and the second item is a regular expression to match hashes (without #). When both match, the hash is allowed.

export type AnchorAllow = [url: RegExp, anchor: RegExp]

Configuration (TypeScript type).

• anchorAllowlist (Array<AnchorAllow>, default: true) — allow anchors; each tuple is checked to match URLs (origin and path, so without search or hash), and then to match hashes (without #); when both match, the hash is allowed, and no missing-anchor error is used
• checkAnchor (boolean, default: true) — check whether URL hashes point to elements
• followMetaHttpEquiv (boolean, default: true) — follow HTML redirects; a <meta content=0;to http-equiv=refresh> can be useful for static sites such as those on GH pages
• findUrls (boolean, default: true) — find URLs in the final resource; currently applies to HTML
• maxRedirects (number, default: 5) — inclusive maximum redirects to follow
• maxRetries (number, default: 1) — inclusive maximum number to try again on failures
• resolveClobberPrefix (boolean, default: true) — accept user-content- prefix in id on elements
• sleep (Sleep, default: defaultSleep) — calculate milliseconds to sleep between tries
• timeout (number, default: 3000) — timeout for HTTP request in milliseconds
• userAgent (string, default: 'Mozilla/5.0 … Safari/537.36', a modern Chrome on macOS user agent) — user agent

Result (TypeScript type).

• messages (Array<VFileMessage>) — messages where the first is a fatal error when dead
• permanent (boolean or undefined) — whether all redirects were permanent
• status ('alive' or 'dead') — status
• url (string or undefined) — final URL if alive
• urls (Set<string> or undefined) — further URLs if findUrls: true and the resource was HTML

Calculate milliseconds to sleep between tries (TypeScript type).

• retries (number) — try

Milliseconds to sleep (number).

Unexpected not ok response `$status` (`$statusText`) on `$url`

This error is used when a URL looks temporarily or permenantly dead.

Unexpected error fetching `$url`

This error is used when for some unknown reason the URL is dead. It might be that you are offline, or the URL is temporarily dead, or something else is wrong. But it looks dead.

Unexpected hash in URL `$url` that redirects with `meta[http-equiv=refresh]` to `$url` losing the hash, remove the hash from the original URL

This warning is used when a URL with an anchor gets an HTML redirect, which looses that anchor. You can remove the hash from the original URL. Or find a similar section on the final URL.

Unexpected hash in URL `$url` to non-html ($contentType) losing the hash, remove the hash from the original URL

This warning is used when a URL with an anchor resolves to some non-HTML thing, which might be an error. Perhaps you can remove the hash from the original URL?

Unexpected hash in URL `$url` that redirects to `$url` losing the hash, remove the hash from the original URL

This warning is used when a URL with an anchor gets an HTTP redirect, which looses that anchor. You can remove the hash from the original URL. Or find a similar section on the final URL.

Unexpected redirect to `$url`, too many redirects

This error is used when there are more redirects than options.maxRedirects. Increase this number or use your browser to see if the URL ends up redirecting to something that works.

Unexpected missing anchor element on `$url` for fragment `$fragment`, remove if unneeded or refer to an existing element

This error is used when a hash is used in a URL but no such element can be found. Elements can match hashes by their id or by their name (when <a>). The prefix user-content- is ignored on elements as that’s what GitHub and rehype-sanitize use to solve DOM clobbering.

Remove the hash if unneeded or refer to a different, existing element.

Unexpected invalid URL `$url` in `content` on `meta[http-equiv=refresh] relative to `$url`

This error is used when an HTML redirect such as <meta content=0;xxx http-equiv=refresh> is used, but xxx does not parse as a URL. Use your browser to see if the URL ends up redirecting to something that works.

This projects is compatible with maintained versions of Node.js.

When we cut a new major release, we drop support for unmaintained versions of Node. This means we try to keep the current release line, dead-or-alive@1, compatible with Node.js 18.

This package is safe.

Yes please! See How to Contribute to Open Source.

MIT © Titus Wormer