browsers.page - Utilities | RTEdge.net docs

Syntax

plain

https://browsers.page/#<options>
https://api.browsers.page/?<parameters>

browsers.page is an RTCode deployment. It reads its options from the URL hash, and any change of the hash reloads the page. api.browsers.page answers GET requests with JSON.

Parameters

The page: hash options

b=<browser[:window]>: A browser to show, with an optional release window. You can repeat it.
B=<group>: A browser group. It is copied to the E parameter of the API.
V=<window>: The release window.
b- and b+: Remove a browser from the list, or add one to it.
css=<CSS>: CSS that the page injects into a <style> element.
n= and v=: Override the detected browser name and version that the page uses to highlight your browser.

Without b and B, the page uses V=[-1,+0] and B=most_used, and its heading reads "Latest" when it is the top-level page. The card of your browser is green when your version is among the releases shown, or newer, and red when it is older than the oldest release shown. The page marks the matching release, or the nearest end of the list.

The API

V=[a,z]: The default release window. The default is [-0,+0], the current release only. + signs are ignored.
b=<id>[:[a,z]]: An explicit browser, with an optional window for that browser. You can repeat it.
b-=<id>: Removes a browser.
b+=<id>: Adds a browser. The key is the literal b+, written b%2B, because a raw b+ decodes to b with a space.
E=desktop|mobile|most_used|runtime: The group to use when no b is given. The default is every available browser.
Q=1: Adds the available member to the answer.

Releases with a release_date are sorted by date, and a window is taken relative to the release whose status is current. Positive offsets therefore reach the beta and nightly releases.

Group	Browsers
`desktop`	`chrome`, `edge`, `firefox`, `opera`, `safari`
`mobile`	`chrome_android`, `firefox_android`, `opera_android`, `safari_ios`, `samsunginternet_android`
`most_used`	`chrome`, `edge`, `firefox`, `safari`, `samsunginternet_android`
`runtime`	`deno`, `nodejs`

On 2026-10-04 the API knew 17 browser ids: bun, chrome, chrome_android, deno, edge, firefox, firefox_android, ie, nodejs, oculus, opera, opera_android, safari, safari_ios, samsunginternet_android, webview_android and webview_ios. An unknown id answers 500 with an error page, not JSON.

Response

The API answers with this shape:

{
  browsers: {
    "<id>": {
      name, type, preview_name, pref_url, accepts_flags, accepts_webextensions,
      releases: { "<version>": { release_date, release_notes, status, engine, engine_version } },
      update_url, current,
      matching: { releases, oldest, newest }
    }
  },
  last_update,
  available
}

preview_name, pref_url, release_notes, update_url and available are present only when there is a value for them. available needs Q=1.
Versions drop a trailing .0, so 15.0 reads as 15.
available is { all, legacy: { desktop: ["ie"] }, evergreen: {…} }.
The headers are Content-Type: application/json;charset=UTF-8, Access-Control-Allow-Origin set to the request's Origin or to *, and Cache-Control: max-age=120.

Examples

A supported-browser table and its data

The iframe shows the desktop browsers with the previous and the current release. The script reads the current version of each desktop browser from the API.

html

<iframe title="Latest browsers" src="//browsers.page#B=desktop&V=[-1,+0]&css=:root{overflow:hidden}" loading=lazy style="inline-size: 100%; block-size: 12rem; border: 0"></iframe>
<ul id=latest></ul>
<script type=module>
  const { browsers } = await fetch(`//api.browsers.page/?${new URLSearchParams({ E: 'desktop' })}`).then(r => r.json());
  latest.innerHTML = Object.values(browsers).map(({ name, current }) => `<li>${name} ${current}</li>`).join('');
</script>

CORS and limits

The API sets Access-Control-Allow-Origin to the request's Origin, or to *, so a page can read it with fetch().
The data comes from MDN's browser compatibility data and the GitHub API, which gives the browser list and the date of the last update.
The worker keeps each response in the Cloudflare cache and returns a stored copy when one matches. It caches upstream fetches for 20 seconds when they succeed, for 1 second on a 404, and not at all on a 5xx. Answers carry Cache-Control: max-age=120.
Embed the page in an iframe, as elefunc.com does, and pass css= in the hash to match your page, for example css=:root{overflow:hidden;color-scheme:light dark}.

Status

The page and the API are available. Their behavior was checked on 2026-10-04.

`browsers.page`