Wikipedia:AT Detector

AT Detector
AT Detector
Release	February 21, 2026; 6 months ago
Written in	JavaScript, React
Platform	Web browser
Type	Wikipedia tool
License	CC0 (public domain)
Website	fixarchive.toolforge.org
Repository	github.com/nethahussain/wikipedia-archive-today-detector

AT Detector is a browser-based reference audit tool that helps editors find and replace archive.today links in Wikipedia articles following the February 2026 consensus to deprecate archive.today. It scans articles for archive.today URLs across all seven known domains, queries the Wayback Machine for replacement snapshots, and provides a guided editing workflow that complies with WP:MEATBOT.

How it works

AT Detector fetches article wikitext via the MediaWiki API and uses a brace-counting state machine to isolate citation templates. It then matches archive.today URLs (including archive.ph, archive.is, archive.fo, archive.li, archive.vn, and archive.md), extracts the embedded original URL from each, and queries the Wayback Machine Availability API for an alternative. Results are displayed with per-link verification and one-click editing tools.

Scan modes

Article: Scan a single article by name, with autocomplete.
Category: Scan up to 500 articles from a category. Optionally includes subcategories up to three levels deep. If fewer than five affected articles are found, the tool automatically continues with a new random batch.
Topic Browse: Scan across 15 predefined topic areas (such as Health & Medicine, Life Sciences, History & Archaeology, Politics & Governance, and others), each mapped to multiple Wikipedia categories.
High Traffic: Scan the most-viewed articles from any given month, using the Wikimedia Pageviews API.

Editing workflow

To comply with WP:MEATBOT, the tool enforces a mandatory verification step before any edit can be staged:

Select an article with detected archive.today links.
For each link, click the suggested Wayback Machine URL to verify the snapshot is an adequate replacement.
Wait for the verification countdown, then click Verify to stage the edit.
Optionally, paste a different Wayback Machine URL if the suggested snapshot is unsuitable. The tool will recalculate the |archive-date= automatically.
Once all desired links are verified, click Push Verified Items to Wikipedia to open the source editor with the replacements pre-applied and the edit summary pre-filled per WP:ATODAY.
Review the diff and publish.

Alternatively, editors can use the manual find-and-replace workflow:

Click Fix on Wikipedia for an individual article.
In the source editor toolbar, open Advanced → Search and replace.
Use the Copy old URL and Copy new URL buttons to populate the Find and Replace fields.
Use Copy date to update |archive-date= in the citation template.
Review and publish.

Guidelines

You are still responsible for your edits: While AT Detector is generally reliable, the Wayback Machine snapshot it suggests may not always be the best available version. Always verify the snapshot before publishing. If the suggested snapshot is incomplete or broken, use the inline URL editor to paste a better one, or skip that link.
Do not make edits without reviewing them: The tool pre-fills the edit, but you should always review the diff before publishing. Check that the replacement URL loads correctly and that the |archive-date= is accurate.
Respect rate limits: The tool is rate-limited to approximately 300 ms between Wayback Machine API requests and auto-retries 503 errors up to three times with exponential backoff.

Troubleshooting

"API request failed": This typically occurs when a browser extension (such as an ad blocker) blocks the tool's API requests. Try disabling your ad blocker for the tool's domain, or use a normal (non-incognito) browser window. Incognito mode may block cross-site script loading.
Wayback Machine 503 errors: This is an Internet Archive server-side issue. The tool automatically retries failed requests. If a Wayback link still shows 503 when you visit it, refresh the page — the archived content is still there.
No results found: If a topic or category scan returns zero affected articles, the archive.today links in that area may have already been cleaned up. Try a different topic or use the Category mode with subcategory depth enabled.

Source code

The source code is available under CC0 (public domain dedication):

github.com/nethahussain/wikipedia-archive-today-detector

Bug reports, feature requests, and pull requests are welcome on GitHub.