All Systems Operational

The Calls channel is fully operational again. Outbound and inbound calls have been restored. Thanks for your patience.

The incident has been resolved

Issue is resolved

We identified the issue affecting our dashboards and it has now been resolved. The dashboards are fully operational again.

Some conversations might have experienced timeouts during this time.

Issue has been resolved

Load times are normalized

Claude is working properly again.

Outbound calls placed through Darwin did not connect. Calls were dialed but never reached the recipient, so no conversation took place and no recording was produced. Affected attempts appear in Darwin as "no answer". Inbound calls were not affected. All other channels — WhatsApp, Instagram, Messenger, TikTok, Webchat and SMS — operated normally throughout. Cause: our outbound telephony provider stopped accepting new call attempts. Service has been restored and verified with a successful outbound test call. If you dialed contacts during this window, those attempts were not delivered and can be re-attempted. Results recorded as "no answer" in this period do not reflect the recipient's behaviour.

SMS delivery was failing across all accounts. Notifications sent by SMS — such as conversation assignment alerts — and outbound SMS messages were not delivered. Other notification channels were not affected: in-app, email, WhatsApp and push notifications continued to work normally throughout, and no data was lost.

Claude access restored

An infrastructure issue at one of our cloud providers caused integrations used by AI Workers — appointment scheduling, availability checks, and other connected services — to fail intermittently or return empty results. In some conversations the AI Worker reported an error or answered as if no information was available. Messaging was not affected. The issue is fully resolved. Actions triggered during this window (such as bookings) may not have completed — we recommend reviewing conversations from this period and retrying any pending actions.

AI Worker responses during voice calls were delayed by several seconds per conversation turn for roughly three and a half hours. Calls completed normally with no data or conversation loss — the only impact was slower AI Worker response time within the call. All other channels (WhatsApp, Instagram, Messenger, TikTok, Webchat) were unaffected. Root cause The AI service that generates responses during calls reached a capacity limit with our provider. Each request had to wait and retry before it could be answered, which added several seconds to every reply. We restored normal response times by routing call traffic through a different, higher-capacity region of the provider's infrastructure.

MCP access was restricted temporarily through the Assistant. We've identified the issue and resolved it.

This incident has been solved, and the latency is back to normal

This issue has been solved, moved by mistake to monitoring