Faster page loads + clearer outage messaging
Performance and reliability release. Static pages, scripts, and styles now stream directly from Cloudflare's edge instead of routing through our worker, so the site loads noticeably quicker. We've also added a friendly banner so you know when something's actually down.
Added2
- If the service is briefly overloaded or under maintenance, you'll now see a clear banner at the top of the page instead of a cryptic error or blank screen. The banner is dismissable and disappears once requests start succeeding again.
- Scheduled maintenance windows are now announced upfront with a clear message and an estimated end time, so you know exactly when the service is intentionally down vs. unexpectedly down. During a maintenance window all consumer-facing requests (sign-ups, logins, inbox actions, the public API, billing) respond with a polite maintenance message; any failed Stripe webhooks or inbound emails will be retried by their senders once the window ends.
Improved2
- Page loads, navigation, and asset delivery are faster across the board. Static files (HTML, JS, CSS, fonts, images) are now served directly from Cloudflare's edge cache instead of going through the worker, cutting 5-30ms off every request and scaling to many more concurrent users without slowing down.
- Better handling of transient edge errors. Network blips and rate limits during high-traffic periods no longer surface as raw HTTP error codes in the UI - the app now treats them as a temporary maintenance state and recovers automatically.
Security1
- Tightened the per-IP rate limit on GET /api/inbox/{inboxId}/messages to 5 requests per minute (down from 120). The official UI and SDKs already poll every 60 seconds, so legitimate clients are unaffected; this stops third-party scrapers and rogue plugins from hammering inbox endpoints. If you're building an integration that needs near-realtime delivery, switch to the SSE /stream endpoint on Pro or Business instead of tight polling.