Every release,
in the open.

Two user-visible fixes and one visual overhaul. The auto-resolve toggle now actually holds the incident open when you switch it off (it was leaking through four other recovery paths we missed). The incident drawer now shows the full update timeline, matching the detail page. And the Past Incidents section on your public status page has been rebuilt to match the layout industry-standard status pages use, date headers grouping the incidents of each day, each incident's full update timeline flat-rendered inline, and the day's wall-clock time shown in your organization's timezone.

• FixAuto-resolve toggle actually works now. Flipping it off prevents every recovery path from closing the incident (previously four of them leaked).
• FixThe incident drawer on the Incidents list now shows the full update timeline (same as the detail page). Previously it only rendered an empty section because the underlying API never returned the updates.
• ImprovedPast Incidents on your public status page is redesigned to a date-grouped timeline matching the layout used by status.claude.com and similar sites. Each day lists its incidents with the full update log inline.
• ImprovedEvery automatic resolution (slow response recovered, SSL renewed, PageSpeed score improved, 1-hour stale cleanup, orphan cleanup) now leaves an explicit System entry in the incident timeline explaining why.

Two changes to how Visual Sentinel closes incidents when a monitor recovers. First, auto-resolutions now leave an explicit 'System' entry in the incident's Updates Timeline so the dashboard and public status page show what happened instead of silently flipping the status. Second, every incident now has an Auto-resolve toggle in the drawer and detail page. Leave it ON (default) for the pipeline to close the incident automatically when the monitor is healthy again. Flip it OFF for incidents where you want to keep the incident open until a team member posts a Resolved update themselves, e.g. while you finish writing the postmortem.

• NewNew Auto-resolve toggle on every incident. ON by default (existing behavior). Flip to OFF to keep the incident open until you manually close it, even if the monitor recovers.
• ImprovedEvery automatic incident resolution now writes a 'System' entry to the update timeline noting that the monitor is back online, so your dashboard and public status page show a clear audit trail instead of a silent status flip.

Three fixes to how your public status page renders incident updates. Your full update timeline now shows on every active and past incident (previously newer updates would push older ones off-screen once you had more than three). Each update now renders with a bold status tag (Investigating, Identified, Monitoring, Resolved, Postmortem) before the message, matching the layout used by the incident-status pages your viewers are used to. Timestamps now display in your organization's timezone instead of the viewer's browser timezone, with the zone abbreviation shown so international viewers can orient themselves.

• FixEvery update on every incident (active and resolved) now shows on your public status page. Previously only the latest three updates per active incident were visible, and resolved incidents showed no updates at all.
• ImprovedEach update now renders with a bold status label (Investigating / Identified / Monitoring / Resolved / Postmortem) before the message, matching standard status-page conventions.
• ImprovedPast Incidents section is now expandable. Click any past incident to drop down its full update timeline, including the final postmortem or resolution note.
• ImprovedTimestamps on the public status page now use your organization's timezone from Settings, with the timezone abbreviation shown next to each time so viewers know what zone they're reading.

The new Post Update form now also appears inside the incident drawer that opens from the Incidents list, so you can post a status update without clicking through to the full details page. We also restyled the form with color-coded status pills (Investigating, Identified, Monitoring, Resolved, Postmortem) matching the badges used elsewhere, an integrated character counter in the message field, and tighter padding that fits the drawer properly.

• ImprovedPost an update directly from the incident drawer on the Incidents list, no need to navigate to the full details page first.
• ImprovedStatus is now a set of color-coded pills (one per lifecycle stage) instead of a plain dropdown, matching the badge styling used across the dashboard.
• ImprovedCharacter counter is integrated into the message field and turns red if you exceed the 2,000-character limit.

Lifecycle emails (Day 1 onboarding, Day 3 visual-monitoring nudge, Day 14 NPS, re-engagement, weekly summary, milestones, monthly product update) are sending again after a delivery regression in a recent release. Transactional mail (alerts, incidents, billing, password reset) was not affected. Lifecycle sends resume on the next scheduled run.

• FixLifecycle emails are back. If you signed up recently and have not received our Day-1 or Day-3 onboarding notes, that is this bug. Emails still due (NPS, milestones, weekly summary) will land on their next scheduled send.
• SecurityOne-click unsubscribe headers (List-Unsubscribe / List-Unsubscribe-Post) are now delivered to your mailbox provider as designed, so the native "Unsubscribe" button in Gmail and Outlook works without a login round-trip.

You can now post status updates directly on any incident and they appear on your public status page in real time. Choose from Investigating, Identified, Monitoring, Resolved, or Postmortem, write a short note for your users, and hit Post. The incident's top-line status moves with it, so your status page tells a clear story from 'we noticed something' through to 'all resolved'. Anyone with Admin or Owner role in the organization can post updates.

• NewNew 'Post an update' card on every incident detail page (Admin or Owner role required). Pick a status, write a message (up to 2,000 characters), and it's live on your public status page immediately.
• NewDashboard incident timeline now shows who posted each update and when, alongside the status label and the message. Posts made automatically by the monitoring pipeline show up as System.
• ImprovedPublic status page continues to show updates chronologically without exposing which staff member posted, keeping the internal and public boundary clean.

Previously only the Owner could save changes to the Organization Settings form, even though Admins could already swap the logo through a separate flow. That was an inconsistency. Admins on the Organization Settings page can now save the form (name, slug, description, website, timezone, billing email, logos) and update whitelabel branding (colors, fonts, email footer, 'Powered by' toggle). Deleting the entire organization and changing outbound SMTP credentials still require the Owner role, because both are either destructive or security-sensitive and not something an Admin should be able to change unilaterally.

• FixAdmins can now save the Organization Settings form. Previously Admins saw an 'Insufficient permissions. Required role: OWNER' banner on Save, while still being able to upload logos separately through a different flow. Both paths are now Admin-accessible.
• ImprovedAdmins can also update whitelabel branding (primary/secondary/accent colors, font family, email footer, the Powered by toggle) and reset branding to defaults. Business plan or higher is still required for the whitelabel feature itself.
• SecurityOrganization deletion and outbound SMTP credentials remain restricted to the Owner role. Both touch destructive or security-sensitive territory where a single-step admin change would be inappropriate.

Monitor Settings → Content Checking lets you flag a monitor as DOWN when required keywords (mustContain) are missing or forbidden keywords (mustNotContain) appear in the response body. The UI accepted your rules and saved them correctly, but the HTTP check pipeline was reading them from the wrong place, so the rules were silently never evaluated. The toggle now works as described. If your saved rules would have been violated by your site's current response, your monitor will correctly flip to a failure state on its next check.

• FixContent Checking rules (mustContain / mustNotContain keywords) are now actually evaluated on every HTTP check. If your site response has ever violated the rules you configured, the monitor will correctly register a failure on the next check (it was silently ignoring violations before). Visual-monitoring keyword checks on the rendered page were unaffected and continue to work as before.

Saving a monitor's Reports settings (include charts, include incidents) was silently failing for some accounts. The save now persists correctly with existing defaults preserved.

• FixMonitor Settings → Reports accordion now saves correctly. The 'Include charts' and 'Include incidents' toggles were sending the right data but failing to persist; defaults preserved so existing monitors behave exactly the same.

A complete rework of which emails you get from us and when. Six distinct drip and engagement flows, all gated by a single one-click unsubscribe link in every footer. Transactional mail (trial reminders, alerts, billing, password reset) stays on separate settings and is never affected. If a flow isn't useful for you, the unsubscribe link flips a flag and every lifecycle email stops.

• NewDay 1 onboarding nudge: if you haven't added a monitor yet, one short email walks you through the 60-second setup. Skipped entirely if you're already monitoring something.
• NewDay 3 visual-monitoring nudge: if you have a monitor but haven't turned on visual monitoring, one email explains what it catches and links to the setting.
• NewDay 14 honest check-in: two weeks in, one email asks whether Visual Sentinel is earning its keep and (if TrustPilot is live for the brand) invites a short review. Detractor path is just 'reply to this email'. The inbox is watched.
• NewDormant re-engagement at 2-3 weeks: if you signed up but never added a monitor, one last soft nudge asking what blocked you. If we can help, reply. If not, the one-click unsubscribe is in the footer.
• NewLapsed re-engagement at 30 days: if you haven't logged in for a month but have monitors running, you get a 30-day activity recap (checks run, incidents detected, uptime, next SSL expiry). No pitch, just the numbers.
• NewFree-tier weekly summary (Monday mornings): seven-day rollup with a context-aware upgrade pitch picked by the engine, not a generic "upgrade now" banner. Incident this week? WhatsApp pitch. Never tried visual monitoring? Visual pitch. Slow site? Performance pitch. Everything in order? No pitch.
• NewMilestones: emails when you catch your first incident, cross 100 and 1,000 checks, and hit your 30-day anniversary. Small dopamine hits with actionable next steps.
• NewMonthly product update digest (1st of each month): summary of what shipped in the last 30 days, pulled directly from this changelog. Skipped automatically if nothing new shipped (no empty digests).
• ImprovedEvery lifecycle email has a one-click unsubscribe link in the footer AND an RFC 8058 header so Gmail and Outlook show their native "Unsubscribe" button right in the inbox. No login required, one click flips a flag server-side.

Sign up with Google, start your 7-day trial, get every feature unlocked. That was already the promise, but a cluster of gating bugs was making the product look like a Free plan with some upsells sprinkled in. This release cleans that up end to end (monitor tabs, schedule step, sidebar badge, billing copy, integrations page).

• FixMonitor detail pages now open Performance, Visual, Content, and DNS History tabs during the trial instead of showing "Unlock with Free Trial" to users who are already on the trial.
• FixSchedule step on the New Monitor wizard now offers 1-minute check frequency during the trial (was capped at 5 minutes).
• FixStatus page limit during the trial now shows the correct allowance instead of being capped at 1.
• FixSidebar plan badge now shows "Trial · Xd" in green during the trial, so you can see your remaining days at a glance. It used to just say "Free".
• FixThe "You're on the Free plan" top banner no longer shows to trial users. It was misleading because the trial already unlocks visual monitoring, WhatsApp/Slack alerts, and everything else on the banner.
• FixBilling page no longer says "billing starts on DATE" for the no-card trial. The no-card trial is truly free: your account simply drops back to the Free plan after 7 days (we email you a reminder before and a recap after, nothing is charged).
• FixCloudflare integration card no longer shows a phantom "Connected · 0 zones" state to users who never linked their Cloudflare account.
• FixAI Insights page loads again instead of erroring out. A stale database field reference was making the page crash on first load.

Narrowing the unified-auth rollout: API keys can no longer reach AI-powered endpoints. Your AI chat assistant, AI insights, incident summarizer, and monitoring digest stay behind your browser session only. This is about preventing token-cost abuse from a leaked key, not a feature regression. Your existing integration work with API keys is unchanged for everything else (monitors, incidents, status pages, teams, etc.).

• SecurityAI endpoints (/api/chat/*, /api/ai/insights, /api/incidents/[id]/summarize, /api/reports/digest) now require a cookie-based session. API keys cannot invoke these features. Reason: AI calls cost money per request and are attractive targets for abuse with a leaked key.
• ImprovedPre-login AI chat widget (/api/chat/public) is unaffected, still publicly available and rate-limited per IP.
• FixInternal build reliability fix bundled with the AI carve-out.

The unified-auth rollout that started in 2.24.0 is now complete. Every tenant-facing API endpoint accepts both session cookies and Bearer API keys. Six endpoints stay session-only by design (auth flows, billing checkout, OAuth init, invitation-accept, account-delete, org-switch). The API docs at /docs now reflect the final state with a dedicated mobile-app section.

• NewBatch 5 migrated 35 more endpoints: all integrations (Discord, Slack, PagerDuty, Webhooks, Google Analytics), chat and support, notification channels read, team member lists, server alerts, AI usage billing, email-tracking stats, report generation, monitor-URL test, incident summaries, and more. API keys with the matching scope can now hit them.
• NewAPI docs at /docs refreshed with a dedicated Mobile Apps section, the full list of session-only exceptions, per-key rate-limit documentation, and a clearer bearer-token authentication walkthrough.
• ImprovedSix endpoints kept session-only by design: auth flows, billing checkout (supports anonymous signup), Slack OAuth init, organization invitation-accept (email-link flow), account deletion (blast-radius guardrail), and organization-switching (cookie mutation).
• ImprovedExisting API keys keep working unchanged. Wildcard (*) scope and matching resource scopes (monitors, channels, organization, analytics, maintenance, user, etc.) now reach every unified endpoint.

Fourth batch of the unified-auth rollout. API keys can now drive onboarding flows (auto-setup, save-step, select-plan, start-no-card-trial, complete, status) and track analytics events (journeys, subscription lifecycle). The API key scope picker now exposes three new resources in the UI: Analytics, Maintenance Windows, and a clearer description for User Profile.

• NewAPI keys can now run the onboarding endpoints (/api/onboarding/auto-setup, /complete, /save-step, /select-plan, /start-no-card-trial, /status) and the analytics endpoints (/api/analytics/journeys, /api/analytics/subscription-events).
• NewAPI key scope picker in Settings -> API Keys now lists three new resources: Analytics (journey + subscription-event tracking), Maintenance Windows (scheduled maintenance), and the User Profile entry description now mentions onboarding.
• ImprovedExisting API keys keep working unchanged. A key with the matching resource scope (analytics, maintenance, user) can hit the newly unified endpoints immediately.
• SecurityAccount deletion (/api/user/delete-account) stays session-only by design. API keys cannot delete the account they belong to.

Third batch of the unified-auth rollout. API keys can now list and read your status pages, read their details, list and read maintenance windows, fetch AI-generated monitoring digests, and check whether all alerts are paused. Your existing keys gain access immediately based on the scopes they already have.

• NewAPI keys can now list and read status pages (/api/status-pages and /api/status-pages/[id]), list and read maintenance windows (/api/maintenance and /api/maintenance/[id]), fetch AI monitoring digests (/api/reports/digest), and check the current alerts-paused state (/api/alerts/pause-all GET).
• ImprovedReport endpoints now map to the "monitors" scope, since reports are derived from monitor data. A key with monitors:read can fetch your digests.
• ImprovedExisting API keys keep working unchanged. A key with a wildcard (*) scope, or the matching resource scope (status-pages, maintenance, monitors, alerts), can hit the newly unified endpoints immediately.

Second batch of the unified-auth rollout. Your API keys can now read your bell-icon notifications, your notification preferences (quiet hours, digest cadence), your digest settings, upload or remove your avatar, and read your current plan + limits + usage. Mobile apps and integrations no longer have to fall back to cookie-based auth for these.

• NewAPI keys can now reach /api/notifications (bell-icon inbox), /api/notification-preferences (quiet hours, grouping, digest), /api/user/digest-preferences, /api/user/profile/avatar (upload + delete), and /api/plan.
• ImprovedAPI key scope mapping tightened: push-subscription endpoints now correctly map to "channels" scope (was falling through to a generic default), plan endpoint maps to "organization" scope, and notification-preferences map to "user" scope.
• ImprovedExisting API keys keep working unchanged. A key with a wildcard (*) scope, or the matching resource scope, can hit the newly unified endpoints immediately.

The first batch of routes that were previously session-only now accept Bearer API keys too. That means integrations can now read your organization, teams, user profile, SMTP settings, and more using the same API key they use for monitor data. Admin endpoints (platform-level) stay session-only. Two new scopes are available: "user" and expanded "organization".

• NewAPI keys can now reach 8 admin endpoints that were previously session-only, including /api/user/profile, /api/organization, /api/teams, /api/organization/smtp, and /api/api-keys list.
• NewNew "user" scope for API keys (read/write) gives access to your profile and preferences without exposing organization-wide data.
• ImprovedThe "organization" scope description in the API key creation dialog now clearly lists what it covers: org settings, teams, members, SMTP, and API-key management.
• ImprovedExisting API keys keep working exactly as before. Integrations with a wildcard (*) scope key, or an "organization" or "user" scoped key, automatically gain access to the newly unified endpoints.

Fixed a subtle bug that affected anyone belonging to multiple organizations. API keys are now locked to the organization you were switched to when you created them, instead of silently resolving to an older org in your account. If you create a key while in Org A, it will always act on Org A data.

• SecurityAPI keys are now tied to the organization they were created in. Previously, if you belonged to multiple organizations, a new key could silently authenticate against a different (usually older) organization than the one you were switched to when you created it.
• FixThe API Keys list page now only shows keys that belong to the currently selected organization. Before, a user in multiple organizations would see the same keys appear under every organization they were a member of.
• ImprovedExisting keys continue to work exactly as they did before, resolving to the same organization they already resolved to. No action required from integrators.

Small SEO polish. The homepage meta description was a few characters over the ideal length used by search engines, so it was getting flagged by Bing. Shortened from 176 to 157 characters.

• ImprovedHomepage meta description shortened to 157 characters, within the 150-160 range recommended by Bing and Google for full display in search results.

Three real visual bugs squashed, plus a silent email bug fix. Your monitor charts no longer dive to 50% uptime on the rightmost data point (that was a partially-sampled current bucket, not a real outage). Y-axis labels no longer get their leading digit clipped on the left. And the weekly / monthly report emails now show a real average response time computed from your checks, not 0ms.

• FixThe "Uptime & Response Time" chart no longer shows a cliff at the rightmost point. The currently-filling 15-minute bucket is now excluded until it has a full sample, so you can stop squinting at "is my site actually 50% down or is that just the chart?".
• FixY-axis labels on every chart across the app no longer get their leading digit clipped (labels like "3600ms" that read as "6600ms"). Affected the overview page, monitor detail, performance tab, server metrics, incident root-cause timeline, and API monitor views.
• FixWeekly organization digest emails and monthly report emails now compute the real average response time from your checks, instead of always showing 0ms. If you have ever looked at those emails and wondered why response time was always zero, this is why.
• ImprovedUnused database column on every monitor dropped. Zero user-visible impact, the column was never populated, just cleanup.