Every release,
in the open.

• FixIf the "site is clear again" message is temporarily held back, for example during a maintenance window or an alert quiet period, it is now retried rather than discarded, so you still hear that a flagged site recovered.

• FixIf the "site is clear again" message cannot be delivered on the first attempt, it is now kept and retried instead of being dropped, so you still find out when a flagged site recovers.

• ImprovedThe "site is clear again" message is now sent only after a site has stayed clean for a short confirmation period, rather than the moment the warning lifts. If the suspicious verification screen comes back during that period, the message is cancelled instead of sent.

• FixThe "site is clear again" message is now held back reliably whenever a suspicious verification screen is found again first, no matter how checks overlap or which internal service is briefly slow.

• FixAlerts about a suspicious verification screen are still delivered if an internal coordination service is briefly unavailable, instead of being held back.

• FixA "site is clear again" message can no longer be sent when a suspicious verification screen is found at the same moment, even when many checks for one monitor overlap.

• FixAlerts about a suspicious verification screen and the matching "site is clear again" message are now handled one at a time per monitor, so overlapping checks cannot deliver them out of order.

• FixThe "site is clear again" message is now held back if the suspicious verification screen reappears at the same moment the warning is being lifted, so you are never told a site is clean while it is still being found.
• FixTaking a screenshot on demand now always returns a result for monitors that were previously flagged.

• FixA site flagged for a suspicious verification screen now only reports as clear when checks running at the same time all agree. Overlapping checks can no longer combine to clear the warning while the problem is still being found.

Page speed checks now retry when the measurement service has a hiccup, and tell you when a result could not be collected.

• FixPage speed checks now retry automatically when the measurement service fails temporarily, so a one-off hiccup no longer leaves a monitor without Core Web Vitals data for the day.
• ImprovedWhen a page speed result cannot be collected, the reason and how long it has been failing are now recorded instead of the result quietly staying blank.
• SecurityEnhanced security hardening.

A site flagged for a suspicious verification screen now clears itself once it is clean again, sub-pages tell you when they are back to normal, and large sites are no longer reported as down by mistake.

• NewWhen a monitor is flagged for a suspicious verification screen, the warning now clears on its own after several clean checks in a row, and you get an alert telling you the site is clear again. Previously the warning stayed until you manually accepted a new baseline.
• ImprovedThe warning on the monitor page now shows whether the problem is still being seen or has not appeared for a while, so you can tell an active issue from an old one.
• FixSub-pages now send a "back to normal" alert when their appearance returns to the baseline, matching what the main page has always done.
• FixSites that send unusually large amounts of header information are no longer reported as down. One monitored site was affected repeatedly.
• FixWhen a first screenshot cannot be taken during setup, the reason shown now describes what actually happened with the site, such as a slow response or a certificate problem, instead of a generic failure message.

Visual baselines now remain stable after capture-setting changes, and every monitored page keeps its own alert state.

• FixAccepting a visual baseline now preserves the monitor’s current viewport and full-page configuration, preventing the next scheduled capture from silently replacing the reviewed baseline.
• FixMain-page and sub-page visual alerts now use independent notification state, so an earlier change on one page cannot suppress a new Slack alert for another page.
• ImprovedSub-page visual alerts and restoration notices now clearly identify the affected page and its URL.
• SecurityEnhanced security hardening.

Monitor details and fleet cards now make health, protection, and report evidence faster to understand on desktop and mobile.

• NewEvery monitor detail page can now generate, track, open, and download a continuous PDF report for the selected 7, 30, or 90-day period.
• ImprovedThe nine primary monitor signals now use consistent cards, plain-language values, visible explanations, and specific health summaries instead of ambiguous status wording.
• ImprovedMonitor fleet cards now show 24-hour uptime and response, SSL and domain expiry, visual-monitoring coverage, check freshness, regions, and clearly labeled actions in a more compact responsive layout.
• FixManual visual captures now consistently use the monitor’s saved viewport and full-page settings unless the request explicitly overrides them.
• SecurityEnhanced security hardening.

The homepage now makes the product easier to evaluate, while monitor details use clearer current-state language and controls.

• ImprovedThe homepage now puts the customer problem, free-plan offer, and interactive product evidence together in a clearer first screen. Its visual comparison and MCP example switch without resizing, and capture statuses remain fully readable.
• ImprovedMonitor details now put HTTP, uptime, response, SSL, domain, DNS, malware-warning, visual, and content status into one balanced summary that links directly to the supporting evidence, uses consistent setup labels, and clearly separates current health from recent trends.
• ImprovedHistorical visual and content changes are now clearly separated from alerts that still need review.
• ImprovedAdvanced monitor sections now expand from their full header, with clearer 24-hour health labels and keyboard-friendly controls.
• ImprovedManual visual captures now show clear in-progress feedback while the latest screenshot is being prepared.

Reviewed captures can once again be accepted cleanly after a website security warning.

• FixFixed an error that could prevent a newer reviewed capture from becoming the visual baseline and clearing its security warning.

Recovery validation now waits correctly while a new snapshot is still being finalized.

• FixAn in-progress recovery snapshot is now allowed to finish before its integrity and freshness are evaluated.

Production recovery now uses frequent verified snapshots and independent freshness monitoring.

• ImprovedRecovery snapshots now protect application data, uploads, configuration, and deployable service artifacts on a twice-hourly schedule.
• ImprovedEvery snapshot is checked after transfer, with independent alerts when recovery coverage becomes stale or unreadable.

Visual checks now protect baseline history when a monitored site is replaced by a suspicious verification screen.

• NewHigh-confidence deceptive verification screens are now quarantined before a screenshot or baseline is saved, with an alert sent through linked visual-notification channels.
• ImprovedMonitor details now show clear safety and cleanup guidance, and the warning remains until a newer reviewed capture is accepted as the baseline.
• SecurityEnhanced security hardening.

Overlapping server reports now preserve the existing alert without generating duplicate operational noise.

• FixServer monitoring now handles an already-open threshold alert without emitting an internal error or dispatching duplicate notifications.
• FixOperational errors that already reached Discord are no longer mirrored there a second time.
• SecurityEnhanced security hardening.

API monitoring pages now show the current entry price and the right check frequency for each plan.

• FixCorrected API monitoring pricing, check intervals, and comparison details across public guidance.