LinkSolva - Broken Link Checker

LinkSolva - Broken Link Checker is a local WordPress tool for administrators who need to know what is genuinely broken before changing content. It discovers URLs in public posts, pages, public custom post types, approved comments, custom navigation menus, featured images, and optionally scalar custom-field or page-builder data. It stores each unique URL separately from the places where it appears, so the same destination is verified once while all source locations remain available. The plugin classifies results instead of treating every failed request as a dead link:

• Healthy — a successful response was received.
• Redirect — the destination works but reaches another URL.
• Broken — HTTP 404 or 410 provides strong evidence that the destination no longer exists.
• Blocked — the destination rejected automated verification, such as HTTP 401 or 403.
• Rate Limited — the destination returned HTTP 429 and should be checked later.
• Temporary Error — the destination returned a 5xx response.
• SSL Error, DNS Error, Timeout, or Unverified — the scanner could not establish enough evidence to call the URL broken.

Every result includes request history, redirect chain, response time, confidence, explanation, and source locations. Local by design

• No cloud account.
• No link credits or usage limits imposed by a third party.
• No telemetry, advertising, or developer tracking.
• No URLs or site content are sent to a developer service.
• HTTP requests are made from your WordPress server to the destinations found in your content.

Safe repairs Supported post content, approved comment content, and custom menu URLs can be replaced, unlinked, or marked nofollow. Builder metadata is read-only in this release. Each repair stores the complete before and after source, the user, time, operation, and checksums. Undo is blocked when the source changed after the repair. Resource controls Scanning runs in small resumable worker requests. Each discovered source batch is verified before the next source batch is read, and repeated occurrences do not re-queue a settled URL. Only one worker request processes the scan at a time. HTTP request timeouts are capped to the worker's remaining time budget, and URLs left unchecked stay queued for the next request. Network failures are recorded without repeating the same request immediately, so the worker can move on and retry later. The scanner also uses per-resource deduplication, redirect limits, response-size limits, and safe retry scheduling. Manual scanning remains available when WordPress cron is delayed. Privacy This plugin does not create a remote account or transmit data to the developer. The server running WordPress may make requests to external URLs found in your content. The request destination, response status, timing, and limited diagnostic categories are stored in your WordPress database.

1.0.10

• Updated the plugin name, directory slug, translation domain, and matching directory graphics for the WordPress.org resubmission.

1.0.9

• Cleared the scan worker lock when the plugin is deactivated or uninstalled, including when findings are retained.
• Required the relevant WordPress source-edit capability before applying or undoing a repair.
• Reused the initial URL safety decision within a parallel batch to avoid an extra DNS validation for each result.
• Paginated navigation-menu discovery so each worker loads only the current small batch of menu items.

1.0.8

• Stopped issuing an immediate GET request after a HEAD network failure or timeout; the URL is recorded as unverified and retried later so the scan can continue.
• Kept batch transport failures and missing responses from triggering a second sequential network attempt.

1.0.7

• Capped each HTTP request to the remaining worker time budget so slow destinations cannot hold a scan request beyond its configured work window.
• Kept unchecked URLs queued for the next worker request when the time budget expires.
• Serialized worker requests so cron and admin polling cannot issue duplicate checks for the same batch.

1.0.6

• Added bounded parallel URL verification with a configurable 1–8 request limit to reduce scan time on large sites.
• Avoided redundant GET requests for definitive 404 and 410 responses while retaining GET fallback for inconclusive HEAD responses.
• Reindexed replacement URLs immediately, verified them before returning success, and retired orphaned old findings.
• Restored the original occurrence/resource mapping on undo and kept replacement and unlink operations auditable.

1.0.5

• Serialized discovery and verification so each discovered batch is classified before more source URLs are fetched.
• Prevented repeated occurrences from resetting an already verified URL to Queued during the same scan.
• Kept ignored resources out of the verification queue while preserving their current-scan visibility.
• Added progress messages and resilient REST polling so active scans continue updating after transient browser requests fail.

1.0.4

• Prevented a scan from completing while current-scan resources remain queued.
• Normalized legacy nullable queue markers during upgrade.
• Removed the stale empty-table row when live results arrive.

1.0.3

• Added Internal links, External links, and All links filtering on the Findings screen.
• Added separate settings to include or exclude internal and external links from scans.
• Reset rediscovered URLs to Queued until their current verification request finishes.
• Updated individual result rows without replacing the complete table body during scan progress.

1.0.2

• Updated scan progress, counters, filters, and result rows in place to prevent dashboard blinking during active scans.
• Preserved the dashboard shell and filter controls while new findings are added.
• Added bounded GET verification when a server rejects or fails a HEAD request.
• Distinguished URLs waiting in the scan queue from completed unverified results.
• Made every new manual or scheduled scan retry previously inconclusive URLs instead of silently skipping them.

1.0.1

• Fixed admin REST requests on sites using plain permalinks.
• Improved DNS failure classification and nonstandard-port URL resolution.
• Preserved every repeated source occurrence while checking each unique URL once.
• Hardened repair auditing, undo behavior, recheck scheduling, translations, and admin asset loading.
• Added a one-minute bounded worker continuation and refreshed WordPress.org release assets.

1.0.0

• New local-only broken link and image checker.
• Added evidence-based HTTP classifications, URL deduplication, bounded scanning, safe repairs, and conflict-aware undo.